From pgsql-performance-owner@postgresql.org Thu Oct 30 12:44:34 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 76963D1B949 for ; Thu, 30 Oct 2003 16:39:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 88821-01 for ; Thu, 30 Oct 2003 12:38:56 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id 2FEE8D1B58D for ; Thu, 30 Oct 2003 12:38:54 -0400 (AST) Received: (qmail 23901 invoked from network); 30 Oct 2003 16:40:09 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 30 Oct 2003 16:40:09 -0000 Message-ID: <3FA323C2.2060302@trade-india.com> Date: Fri, 31 Oct 2003 22:08:50 -0500 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Christopher Browne Cc: pgsql-performance@postgresql.org Subject: Re: Query puts 7.3.4 on endless loop but 7.4beta5 is fine. References: <33298.203.145.130.142.1067457626.squirrel@mail.trade-india.com> <20031029202920.GZ7337@pallas.fsck.com> <33402.203.145.130.142.1067460429.squirrel@mail.trade-india.com> <200310301702.00275.mallah@trade-india.com> <13042.61.16.154.82.1067523120.squirrel@mail.trade-india.com> In-Reply-To: Content-Type: multipart/alternative; boundary="------------000303000106050002000202" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200310/761 X-Sequence-Number: 4509 This is a multi-part message in MIME format. --------------000303000106050002000202 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Hi , Here are the Execution Plans , Sorry for the delay . Regds Mallah On PostgreSQL 7.3.4 rt3=# explain SELECT DISTINCT main.* FROM ((((Tickets main JOIN Groups as Groups_1 ON ( main.id = Groups_1.Instance)) JOIN Principals as Principals_2 ON ( Groups_1.id = Principals_2.ObjectId)) JOIN CachedGroupMembers as CachedGroupMembers_3 ON ( Principals_2.id = CachedGroupMembers_3.GroupId)) JOIN Users as Users_4 ON ( CachedGroupMembers_3.MemberId = Users_4.id)) WHERE ((main.EffectiveId = main.id)) AND ((main.Type = 'ticket')) AND ( ( ( (Users_4.EmailAddress = 'mallah_rajesh@yahoo.com') AND(Groups_1.Domain = 'RT::Ticket-Role')AND(Groups_1.Type = 'Requestor')AND(Principals_2.PrincipalType = 'Group') ) ) AND ( (main.Status = 'new')OR(main.Status = 'open') ) ) ORDER BY main.Priority DESC LIMIT 10; Limit (cost=2044.52..2044.58 rows=1 width=195) -> Unique (cost=2044.52..2044.58 rows=1 width=195) -> Sort (cost=2044.52..2044.52 rows=1 width=195) Sort Key: main.priority, main.id, main.effectiveid, main.queue, main."type", main.issuestatement, main.resolution, main."owner", main.subject, main.initialpriority, main.finalpriority, main.timeestimated, main.timeworked, main.status, main.timeleft, main.told, main.starts, main.started, main.due, main.resolved, main.lastupdatedby, main.lastupdated, main.creator, main.created, main.disabled -> Hash Join (cost=3.98..2044.51 rows=1 width=195) Hash Cond: ("outer".memberid = "inner".id) -> Nested Loop (cost=0.00..2040.51 rows=2 width=191) -> Nested Loop (cost=0.00..1914.41 rows=1 width=183) -> Nested Loop (cost=0.00..1909.67 rows=1 width=175) Join Filter: (("outer".id)::text = ("inner".instance)::text) -> Seq Scan on tickets main (cost=0.00..465.62 rows=1 width=163) Filter: ((effectiveid = id) AND ("type" = 'ticket'::character varying) AND ((status = 'new'::character varying) OR (status = 'open'::character varying))) -> Index Scan using groups_domain on groups groups_1 (cost=0.00..1338.03 rows=7068 width=12) Index Cond: ("domain" = 'RT::Ticket-Role'::character varying) Filter: ("type" = 'Requestor'::character varying) -> Index Scan using principals2 on principals principals_2 (cost=0.00..4.73 rows=1 width=8) Index Cond: ("outer".id = principals_2.objectid) Filter: (principaltype = 'Group'::character varying) -> Index Scan using cachedgroupmembers3 on cachedgroupmembers cachedgroupmembers_3 (cost=0.00..125.54 rows=45 width=8) Index Cond: ("outer".id = cachedgroupmembers_3.groupid) -> Hash (cost=3.98..3.98 rows=1 width=4) -> Index Scan using users4 on users users_4 (cost=0.00..3.98 rows=1 width=4) Index Cond: (emailaddress = 'mallah_rajesh@yahoo.com'::character varying) (23 rows) On PostgreSQL 7.4 beta 5 rt3=# explain SELECT DISTINCT main.* FROM ((((Tickets main JOIN Groups as Groups_1 ON ( main.id = Groups_1.Instance)) JOIN Principals as Principals_2 ON ( Groups_1.id = Principals_2.ObjectId)) JOIN CachedGroupMembers as CachedGroupMembers_3 ON ( Principals_2.id = CachedGroupMembers_3.GroupId)) JOIN Users as Users_4 ON ( CachedGroupMembers_3.MemberId = Users_4.id)) WHERE ((main.EffectiveId = main.id)) AND ((main.Type = 'ticket')) AND ( ( ( (Users_4.EmailAddress = 'mallah_rajesh@yahoo.com') AND(Groups_1.Domain = 'RT::Ticket-Role')AND(Groups_1.Type = 'Requestor')AND(Principals_2.PrincipalType = 'Group') ) ) AND ( (main.Status = 'new')OR(main.Status = 'open') ) ) ORDER BY main.Priority DESC LIMIT 10; QUERY PLAN --------------------------------------------------------------- Limit (cost=582.27..582.34 rows=1 width=164) -> Unique (cost=582.27..582.34 rows=1 width=164) -> Sort (cost=582.27..582.28 rows=1 width=164) Sort Key: main.priority, main.id, main.effectiveid, main.queue, main."type", main.issuestatement, main.resolution, main."owner", main.subject, main.initialpriority, main.finalpriority, main.timeestimated, main.timeworked, main.status, main.timeleft, main.told, main.starts, main.started, main.due, main.resolved, main.lastupdatedby, main.lastupdated, main.creator, main.created, main.disabled -> Hash Join (cost=476.18..582.26 rows=1 width=164) Hash Cond: ("outer".groupid = "inner".id) -> Nested Loop (cost=0.00..105.97 rows=21 width=4) -> Index Scan using users4 on users users_4 (cost=0.00..3.99 rows=2 width=4) Index Cond: ((emailaddress)::text = 'mallah_rajesh@yahoo.com'::text) -> Index Scan using cachedgroupmembers2 on cachedgroupmembers cachedgroupmembers_3 (cost=0.00..50.81 rows=14 width=8) Index Cond: (cachedgroupmembers_3.memberid = "outer".id) -> Hash (cost=476.17..476.17 rows=1 width=168) -> Nested Loop (cost=0.00..476.17 rows=1 width=168) -> Nested Loop (cost=0.00..471.54 rows=1 width=168) -> Seq Scan on tickets main (cost=0.00..465.62 rows=1 width=164) Filter: ((effectiveid = id) AND (("type")::text = 'ticket'::text) AND (((status)::text = 'new'::text) OR ((status)::text = 'open'::text))) -> Index Scan using groups1 on groups groups_1 (cost=0.00..5.90 rows=1 width=12) Index Cond: (((groups_1."domain")::text = 'RT::Ticket-Role'::text) AND (("outer".id)::text = (groups_1.instance)::text) AND ((groups_1."type")::text = 'Requestor'::text)) -> Index Scan using principals2 on principals principals_2 (cost=0.00..4.62 rows=1 width=8) Index Cond: ("outer".id = principals_2.objectid) Filter: ((principaltype)::text = 'Group'::text) (21 rows) rt3=# Christopher Browne wrote: >In the last exciting episode, mallah@trade-india.com wrote: > > >>>mallah@trade-india.com (Rajesh Kumar Mallah) wrote: >>> >>> >>>>Can you please have a Look at the below and suggest why it >>>>apparently puts 7.3.4 on an infinite loop . the CPU utilisation of the backend running it >>>>approches 99%. >>>> >>>> >>>What would be useful, for this case, would be to provide the query plan, perhaps via >>> >>> EXPLAIN [Big Long Query]. >>> >>>The difference between that EXPLAIN and what you get on 7.4 might be quite interesting. >>> >>>I would think it quite unlikely that it is truly an "infinite" loop; it is rather more likely >>>that the plan winds up being pretty bad and doing something [a bunch of nested loops, maybe?] >>>that run longer than your patience will permit. >>> >>> >>:-) ok i will leave it running and try to get it. >> >> > >No, if you just do EXPLAIN (and not EXPLAIN ANALYZE), that returns >without executing the query. > >If the query runs for a really long time, then we _know_ that there is >something troublesome. EXPLAIN (no ANALYZE) should provide some >insight without having anything run for a long time. > >If EXPLAIN [big long query] turns into what you are terming an >"infinite loop," then you have a quite different problem, and it would >be very useful to know that. > > --------------000303000106050002000202 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit
Hi ,

Here are the Execution Plans ,
Sorry for the delay .

Regds
Mallah
On PostgreSQL  7.3.4 

rt3=# explain   SELECT DISTINCT main.* FROM ((((Tickets main  JOIN Groups as Groups_1  ON ( main.id = Groups_1.Instance))  
JOIN Principals as Principals_2  ON ( Groups_1.id = Principals_2.ObjectId)) JOIN CachedGroupMembers as CachedGroupMembers_3  
ON ( Principals_2.id = CachedGroupMembers_3.GroupId))  JOIN Users as Users_4  ON ( CachedGroupMembers_3.MemberId = Users_4.id))   
WHERE ((main.EffectiveId = main.id)) AND ((main.Type = 'ticket')) AND ( (  ( (Users_4.EmailAddress = 'mallah_rajesh@yahoo.com')
AND(Groups_1.Domain = 'RT::Ticket-Role')AND(Groups_1.Type = 'Requestor')AND(Principals_2.PrincipalType = 'Group') )  ) 
AND ( (main.Status = 'new')OR(main.Status = 'open') ) )  ORDER BY main.Priority DESC LIMIT 10;



Limit  (cost=2044.52..2044.58 rows=1 width=195)
   ->  Unique  (cost=2044.52..2044.58 rows=1 width=195)
         ->  Sort  (cost=2044.52..2044.52 rows=1 width=195)
               Sort Key: main.priority, main.id, main.effectiveid, main.queue, main."type", main.issuestatement, main.resolution, main."owner", main.subject, main.initialpriority, main.finalpriority, main.timeestimated, main.timeworked, main.status, main.timeleft, main.told, main.starts, main.started, main.due, main.resolved, main.lastupdatedby, main.lastupdated, main.creator, main.created, main.disabled
               ->  Hash Join  (cost=3.98..2044.51 rows=1 width=195)
                     Hash Cond: ("outer".memberid = "inner".id)
                     ->  Nested Loop  (cost=0.00..2040.51 rows=2 width=191)
                           ->  Nested Loop  (cost=0.00..1914.41 rows=1 width=183)
                                 ->  Nested Loop  (cost=0.00..1909.67 rows=1 width=175)
                                       Join Filter: (("outer".id)::text = ("inner".instance)::text)
                                       ->  Seq Scan on tickets main  (cost=0.00..465.62 rows=1 width=163)
                                             Filter: ((effectiveid = id) AND ("type" = 'ticket'::character varying) AND ((status = 'new'::character varying) OR (status = 'open'::character varying)))
                                       ->  Index Scan using groups_domain on groups groups_1  (cost=0.00..1338.03 rows=7068 width=12)
                                             Index Cond: ("domain" = 'RT::Ticket-Role'::character varying)
                                             Filter: ("type" = 'Requestor'::character varying)
                                 ->  Index Scan using principals2 on principals principals_2  (cost=0.00..4.73 rows=1 width=8)
                                       Index Cond: ("outer".id = principals_2.objectid)
                                       Filter: (principaltype = 'Group'::character varying)
                           ->  Index Scan using cachedgroupmembers3 on cachedgroupmembers cachedgroupmembers_3  (cost=0.00..125.54 rows=45 width=8)
                                 Index Cond: ("outer".id = cachedgroupmembers_3.groupid)
                     ->  Hash  (cost=3.98..3.98 rows=1 width=4)
                           ->  Index Scan using users4 on users users_4  (cost=0.00..3.98 rows=1 width=4)
                                 Index Cond: (emailaddress = 'mallah_rajesh@yahoo.com'::character varying)
(23 rows)
On PostgreSQL 7.4 beta 5

rt3=# explain   SELECT DISTINCT main.* FROM ((((Tickets main  JOIN Groups as Groups_1  ON ( main.id = Groups_1.Instance))  
JOIN Principals as Principals_2  ON ( Groups_1.id = Principals_2.ObjectId)) JOIN CachedGroupMembers as CachedGroupMembers_3  
ON ( Principals_2.id = CachedGroupMembers_3.GroupId))  JOIN Users as Users_4  ON ( CachedGroupMembers_3.MemberId = Users_4.id))   
WHERE ((main.EffectiveId = main.id)) AND ((main.Type = 'ticket')) AND ( (  ( (Users_4.EmailAddress = 'mallah_rajesh@yahoo.com')
AND(Groups_1.Domain = 'RT::Ticket-Role')AND(Groups_1.Type = 'Requestor')AND(Principals_2.PrincipalType = 'Group') )  ) 
AND ( (main.Status = 'new')OR(main.Status = 'open') ) )  ORDER BY main.Priority DESC LIMIT 10;
                                                                                                                                                                                                       QUERY PLAN
---------------------------------------------------------------
 Limit  (cost=582.27..582.34 rows=1 width=164)
   ->  Unique  (cost=582.27..582.34 rows=1 width=164)
         ->  Sort  (cost=582.27..582.28 rows=1 width=164)
               Sort Key: main.priority, main.id, main.effectiveid, main.queue, main."type", main.issuestatement, main.resolution, main."owner", main.subject, main.initialpriority, main.finalpriority, main.timeestimated, main.timeworked, main.status, main.timeleft, main.told, main.starts, main.started, main.due, main.resolved, main.lastupdatedby, main.lastupdated, main.creator, main.created, main.disabled
               ->  Hash Join  (cost=476.18..582.26 rows=1 width=164)
                     Hash Cond: ("outer".groupid = "inner".id)
                     ->  Nested Loop  (cost=0.00..105.97 rows=21 width=4)
                           ->  Index Scan using users4 on users users_4  (cost=0.00..3.99 rows=2 width=4)
                                 Index Cond: ((emailaddress)::text = 'mallah_rajesh@yahoo.com'::text)
                           ->  Index Scan using cachedgroupmembers2 on cachedgroupmembers cachedgroupmembers_3  (cost=0.00..50.81 rows=14 width=8)
                                 Index Cond: (cachedgroupmembers_3.memberid = "outer".id)
                     ->  Hash  (cost=476.17..476.17 rows=1 width=168)
                           ->  Nested Loop  (cost=0.00..476.17 rows=1 width=168)
                                 ->  Nested Loop  (cost=0.00..471.54 rows=1 width=168)
                                       ->  Seq Scan on tickets main  (cost=0.00..465.62 rows=1 width=164)
                                             Filter: ((effectiveid = id) AND (("type")::text = 'ticket'::text) AND (((status)::text = 'new'::text) OR ((status)::text = 'open'::text)))
                                       ->  Index Scan using groups1 on groups groups_1  (cost=0.00..5.90 rows=1 width=12)
                                             Index Cond: (((groups_1."domain")::text = 'RT::Ticket-Role'::text) AND (("outer".id)::text = (groups_1.instance)::text) AND ((groups_1."type")::text = 'Requestor'::text))
                                 ->  Index Scan using principals2 on principals principals_2  (cost=0.00..4.62 rows=1 width=8)
                                       Index Cond: ("outer".id = principals_2.objectid)
                                       Filter: ((principaltype)::text = 'Group'::text)
(21 rows)

rt3=#
Christopher Browne wrote:
In the last exciting episode, mallah@trade-india.com wrote:
  
mallah@trade-india.com (Rajesh Kumar Mallah) wrote:
      
Can you please have a Look at the below and suggest why it
apparently puts 7.3.4 on an infinite loop . the CPU utilisation of the backend running it
approches 99%.
        
What would be useful, for this case, would be to provide the query plan, perhaps via

 EXPLAIN [Big Long Query].

The difference between that EXPLAIN and what you get on 7.4 might be quite interesting.

I would think it quite unlikely that it is truly an "infinite" loop; it is rather more likely
that the plan winds up being pretty bad and doing something [a bunch of nested loops, maybe?]
that run longer than your patience will permit.
      
:-)   ok i will leave it running and try to get it.
    

No, if you just do EXPLAIN (and not EXPLAIN ANALYZE), that returns
without executing the query.

If the query runs for a really long time, then we _know_ that there is
something troublesome.  EXPLAIN (no ANALYZE) should provide some
insight without having anything run for a long time.

If EXPLAIN [big long query] turns into what you are terming an
"infinite loop," then you have a quite different problem, and it would
be very useful to know that.

--------------000303000106050002000202-- From pgsql-performance-owner@postgresql.org Thu Oct 30 13:04:58 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id BE5A0D1B8D7 for ; Thu, 30 Oct 2003 16:50:26 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 88821-08 for ; Thu, 30 Oct 2003 12:49:56 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id 1AD9BD1B923 for ; Thu, 30 Oct 2003 12:49:54 -0400 (AST) Received: (qmail 24761 invoked from network); 30 Oct 2003 16:51:11 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 30 Oct 2003 16:51:11 -0000 Message-ID: <3FA3265B.6040707@trade-india.com> Date: Fri, 31 Oct 2003 22:19:55 -0500 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Tom Lane Cc: pgsql-performance@postgresql.org Subject: Re: Query puts 7.3.4 on endless loop but 7.4beta5 is fine. References: <33298.203.145.130.142.1067457626.squirrel@mail.trade-india.com> <20031029202920.GZ7337@pallas.fsck.com> <33402.203.145.130.142.1067460429.squirrel@mail.trade-india.com> <200310301702.00275.mallah@trade-india.com> <28551.1067530272@sss.pgh.pa.us> In-Reply-To: <28551.1067530272@sss.pgh.pa.us> Content-Type: multipart/alternative; boundary="------------010705060908090409080604" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200310/763 X-Sequence-Number: 4511 This is a multi-part message in MIME format. --------------010705060908090409080604 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Tom Lane wrote: >Rajesh Kumar Mallah writes: > > >>SELECT DISTINCT main.* FROM >>( >> ( >> ( >> ( >> Tickets main JOIN Groups as Groups_1 ON ( main.id = Groups_1.Instance) >> ) JOIN >> Principals as Principals_2 ON ( Groups_1.id = Principals_2.ObjectId) >> ) JOIN >> CachedGroupMembers as CachedGroupMembers_3 ON ( Principals_2.id = CachedGroupMembers_3.GroupId) >> ) JOIN >> Users as Users_4 ON ( CachedGroupMembers_3.MemberId = Users_4.id) >>) WHERE >> ... >> >> > >I think the reason for the performance difference is that 7.3 treats >JOIN syntax as forcing a particular join order, while 7.4 doesn't. > Just out of curiosity , how does 7.4 determine the optimal Join Order? is it GEQO in case of 7.4 although i did not enable it explicitly? Thanks for the reply , I sent the EXPLAINs also just now. What i really want is to help improving the Pg specific Component for DBIx::SearchBuilder. The module is being widely used in the mod_perl world and has impact on the performance perception of PostgreSQL. > > regards, tom lane > > --------------010705060908090409080604 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Tom Lane wrote:
Rajesh Kumar Mallah <mallah@trade-india.com> writes:
  
SELECT DISTINCT main.* FROM
(
   (
      (
         (
            Tickets main  JOIN Groups as Groups_1  ON ( main.id = Groups_1.Instance)
         )  JOIN
         Principals as Principals_2  ON ( Groups_1.id = Principals_2.ObjectId)
      ) JOIN
      CachedGroupMembers as CachedGroupMembers_3  ON ( Principals_2.id = CachedGroupMembers_3.GroupId)
   )  JOIN
   Users as Users_4  ON ( CachedGroupMembers_3.MemberId = Users_4.id)
)   WHERE
 ...
    

I think the reason for the performance difference is that 7.3 treats
JOIN syntax as forcing a particular join order, while 7.4 doesn't.

Just out of curiosity , how does 7.4 determine the optimal Join Order?
is it GEQO in case of 7.4 although i did not enable it explicitly?
Thanks for the reply , I sent the EXPLAINs also just now.

What i really want is to help improving the Pg specific Component
for DBIx::SearchBuilder. The module is being widely used in
the  mod_perl world and has impact on the performance perception
of PostgreSQL.

			regards, tom lane

--------------010705060908090409080604-- From pgsql-performance-owner@postgresql.org Thu Oct 30 15:17:01 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3D439D1B4E1 for ; Thu, 30 Oct 2003 19:16:59 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 11759-08 for ; Thu, 30 Oct 2003 15:16:29 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id 5C145D1B536 for ; Thu, 30 Oct 2003 15:16:27 -0400 (AST) Received: (qmail 6549 invoked from network); 30 Oct 2003 19:17:36 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 30 Oct 2003 19:17:36 -0000 Message-ID: <3FA34827.1030705@trade-india.com> Date: Sat, 01 Nov 2003 00:44:07 -0500 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Greg Stark Cc: Christopher Browne , pgsql-performance@postgresql.org, Jesse Vincent Subject: Re: Query puts 7.3.4 on endless loop but 7.4beta5 is fine. References: <33298.203.145.130.142.1067457626.squirrel@mail.trade-india.com> <20031029202920.GZ7337@pallas.fsck.com> <33402.203.145.130.142.1067460429.squirrel@mail.trade-india.com> <200310301702.00275.mallah@trade-india.com> <13042.61.16.154.82.1067523120.squirrel@mail.trade-india.com> <3FA323C2.2060302@trade-india.com> <87r80ud0nw.fsf@stark.dyndns.tv> In-Reply-To: <87r80ud0nw.fsf@stark.dyndns.tv> Content-Type: multipart/alternative; boundary="------------070306070203080407090902" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200310/772 X-Sequence-Number: 4520 This is a multi-part message in MIME format. --------------070306070203080407090902 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit explain analyze of original Query: rt3=# explain analyze SELECT DISTINCT main.* FROM Tickets main JOIN Groups as Groups_1 ON ( main.id = Groups_1.Instance) JOIN Principals as Principals_2 ON ( Groups_1.id = Principals_2.ObjectId) JOIN CachedGroupMembers as CachedGroupMembers_3 ON ( Principals_2.id = CachedGroupMembers_3.GroupId) JOIN Users as Users_4 ON ( CachedGroupMembers_3.MemberId = Users_4.id) WHERE ((main.EffectiveId = main.id)) AND ((main.Type = 'ticket')) AND ( ( ( (lower(Users_4.EmailAddress) = 'mallah_rajesh@yahoo.com')AND(Groups_1.Domain = 'RT::Ticket-Role')AND(Groups_1.Type = 'Requestor')AND(Principals_2.PrincipalType = 'Group') ) ) AND ( (main.Status = 'new')OR(main.Status = 'open') ) ) ORDER BY main.Priority DESC LIMIT 10; QUERY PLAN ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ Limit (cost=619.93..620.00 rows=1 width=164) (actual time=994.570..994.683 rows=4 loops=1) -> Unique (cost=619.93..620.00 rows=1 width=164) (actual time=994.565..994.672 rows=4 loops=1) -> Sort (cost=619.93..619.93 rows=1 width=164) (actual time=994.561..994.569 rows=8 loops=1) Sort Key: main.priority, main.id, main.effectiveid, main.queue, main."type", main.issuestatement, main.resolution, main."owner", main.subject, main.initialpriority, main.finalpriority, main.timeestimated, main.timeworked, main.status, main.timeleft, main.told, main.starts, main.started, main.due, main.resolved, main.lastupdatedby, main.lastupdated, main.creator, main.created, main.disabled -> Nested Loop (cost=0.00..619.92 rows=1 width=164) (actual time=1.374..993.998 rows=8 loops=1) -> Nested Loop (cost=0.00..610.83 rows=3 width=168) (actual time=0.691..839.633 rows=9617 loops=1) -> Nested Loop (cost=0.00..476.17 rows=1 width=168) (actual time=0.524..616.937 rows=3209 loops=1) -> Nested Loop (cost=0.00..471.54 rows=1 width=168) (actual time=0.376..503.774 rows=3209 loops=1) -> Seq Scan on tickets main (cost=0.00..465.62 rows=1 width=164) (actual time=0.114..60.044 rows=3209 loops=1) Filter: ((effectiveid = id) AND (("type")::text = 'ticket'::text) AND (((status)::text = 'new'::text) OR ((status)::text = 'open'::text))) -> Index Scan using groups1 on groups groups_1 (cost=0.00..5.90 rows=1 width=12) (actual time=0.111..0.119 rows=1 loops=3209) Index Cond: (((groups_1."domain")::text = 'RT::Ticket-Role'::text) AND (("outer".id)::text = (groups_1.instance)::text) AND ((groups_1."type")::text = 'Requestor'::text)) -> Index Scan using principals2 on principals principals_2 (cost=0.00..4.62 rows=1 width=8) (actual time=0.015..0.018 rows=1 loops=3209) Index Cond: ("outer".id = principals_2.objectid) Filter: ((principaltype)::text = 'Group'::text) -> Index Scan using cachedgroupmembers3 on cachedgroupmembers cachedgroupmembers_3 (cost=0.00..134.06 rows=47 width=8) (actual time=0.015..0.026 rows=3 loops=3209) Index Cond: ("outer".id = cachedgroupmembers_3.groupid) -> Index Scan using users_pkey on users users_4 (cost=0.00..3.02 rows=1 width=4) (actual time=0.013..0.013 rows=0 loops=9617) Index Cond: ("outer".memberid = users_4.id) Filter: (lower((emailaddress)::text) = 'mallah_rajesh@yahoo.com'::text) Total runtime: 995.326 ms (21 rows) rt3=# 999 ms is not that bad but u think it deserves this many ms? Nopes the query are not Equiv , earlier one returns 4 rows and the below one none, can you spot any obvious and resend plz. thats why i did not do an explain analyze rt3=# SELECT * rt3-# FROM tickets rt3-# WHERE id IN ( rt3(# SELECT groups.instance rt3(# FROM groups rt3(# JOIN principals ON (groups.id = principals.objectid) rt3(# JOIN cachedgroupmembers ON (principals.id = cachedgroupmembers.groupid) rt3(# JOIN users ON (cachedgroupmembers.memberid = users.id) rt3(# WHERE users.emailaddress = 'mallah_rajesh@yahoo.com' rt3(# AND groups.domain = 'RT::Ticket-Role' rt3(# AND groups.type = 'Requestor' rt3(# AND principals.principaltype = 'group' rt3(# ) rt3-# AND type = 'ticket' rt3-# AND effectiveid = tickets.id rt3-# AND (status = 'new' OR status = 'open') rt3-# ORDER BY priority DESC rt3-# LIMIT 10; id | effectiveid | queue | type | issuestatement | resolution | owner | subject | initialpriority | finalpriority | priority | timeestimated | timeworked | status | timeleft | told | starts | started | due | resolved | lastupdatedby | lastupdated | creator | created | disabled ----+-------------+-------+------+----------------+------------+-------+---------+-----------------+---------------+----------+---------------+------------+--------+----------+------+--------+---------+-----+----------+---------------+-------------+---------+---------+---------- (0 rows) Time: 2670.85 ms rt3=# Well it may be of interest to write the query in best possible way but i am not sure if it really helps the RT application becoz i do not know whether DBIx::SearchBuilder would currently allow auto generation of such arbitrary SQLs. Regds Mallah. Greg Stark wrote: >Rajesh Kumar Mallah writes: > > > >>rt3=# explain >> >>SELECT DISTINCT main.* >> FROM ((( >> (Tickets main JOIN Groups as Groups_1 ON ( main.id = Groups_1.Instance)) >> JOIN Principals as Principals_2 ON ( Groups_1.id = Principals_2.ObjectId) >> ) JOIN CachedGroupMembers as CachedGroupMembers_3 ON ( Principals_2.id = CachedGroupMembers_3.GroupId) >> ) JOIN Users as Users_4 ON ( CachedGroupMembers_3.MemberId = Users_4.id) >> ) >> WHERE ((main.EffectiveId = main.id)) >> AND ((main.Type = 'ticket')) >> AND ((( (Users_4.EmailAddress = 'mallah_rajesh@yahoo.com') >> AND (Groups_1.Domain = 'RT::Ticket-Role') >> AND (Groups_1.Type = 'Requestor') >> AND (Principals_2.PrincipalType = 'Group') >> )) >> AND ((main.Status = 'new') OR (main.Status = 'open')) >> ) >> ORDER BY main.Priority DESC LIMIT 10; >> >> > >So this query seems to be going the long way around to do the equivalent of an >IN clause. Presumably because as far as I know mysql didn't support IN >subqueries until recently. > >Can you do an "explain analyze" on the above query and the following rewritten >one in 7.4? The "analyze" is important because it'll give real timing >information. And it's important that it be on 7.4 as there were improvements >in this area specifically in 7.4. > >SELECT * > FROM tickets > WHERE id IN ( > SELECT groups.instance > FROM groups > JOIN principals ON (groups.id = principals.objectid) > JOIN cachedgroupmembers ON (principals.id = cachedgroupmembers.groupid) > JOIN users ON (cachedgroupmembers.memberid = users.id) > WHERE users.emailaddress = 'mallah_rajesh@yahoo.com' > AND groups.domain = 'RT::Ticket-Role' > AND groups.type = 'Requestor' > AND principals.principaltype = 'group' > ) > AND type = 'ticket' > AND effectiveid = tickets.id > AND (status = 'new' OR status = 'open') >ORDER BY priority DESC >LIMIT 10; > > > > > > > > > --------------070306070203080407090902 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit

explain analyze of original Query:

rt3=# explain analyze SELECT DISTINCT main.* FROM Tickets main  JOIN Groups as Groups_1  ON ( main.id = Groups_1.Instance)   JOIN Principals as Principals_2  ON ( Groups_1.id = Principals_2.ObjectId)   JOIN CachedGroupMembers as CachedGroupMembers_3  ON ( Principals_2.id = CachedGroupMembers_3.GroupId)   JOIN Users as Users_4  ON ( CachedGroupMembers_3.MemberId = Users_4.id)    WHERE ((main.EffectiveId = main.id)) AND ((main.Type = 'ticket')) AND ( (  ( (lower(Users_4.EmailAddress) = 'mallah_rajesh@yahoo.com')AND(Groups_1.Domain = 'RT::Ticket-Role')AND(Groups_1.Type = 'Requestor')AND(Principals_2.PrincipalType = 'Group') )  ) AND ( (main.Status = 'new')OR(main.Status = 'open') ) )  ORDER BY main.Priority DESC LIMIT 10;
                                                                                                                                                                                                       QUERY PLAN
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 Limit  (cost=619.93..620.00 rows=1 width=164) (actual time=994.570..994.683 rows=4 loops=1)
   ->  Unique  (cost=619.93..620.00 rows=1 width=164) (actual time=994.565..994.672 rows=4 loops=1)
         ->  Sort  (cost=619.93..619.93 rows=1 width=164) (actual time=994.561..994.569 rows=8 loops=1)
               Sort Key: main.priority, main.id, main.effectiveid, main.queue, main."type", main.issuestatement, main.resolution, main."owner", main.subject, main.initialpriority, main.finalpriority, main.timeestimated, main.timeworked, main.status, main.timeleft, main.told, main.starts, main.started, main.due, main.resolved, main.lastupdatedby, main.lastupdated, main.creator, main.created, main.disabled
               ->  Nested Loop  (cost=0.00..619.92 rows=1 width=164) (actual time=1.374..993.998 rows=8 loops=1)
                     ->  Nested Loop  (cost=0.00..610.83 rows=3 width=168) (actual time=0.691..839.633 rows=9617 loops=1)
                           ->  Nested Loop  (cost=0.00..476.17 rows=1 width=168) (actual time=0.524..616.937 rows=3209 loops=1)
                                 ->  Nested Loop  (cost=0.00..471.54 rows=1 width=168) (actual time=0.376..503.774 rows=3209 loops=1)
                                       ->  Seq Scan on tickets main  (cost=0.00..465.62 rows=1 width=164) (actual time=0.114..60.044 rows=3209 loops=1)
                                             Filter: ((effectiveid = id) AND (("type")::text = 'ticket'::text) AND (((status)::text = 'new'::text) OR ((status)::text = 'open'::text)))
                                       ->  Index Scan using groups1 on groups groups_1  (cost=0.00..5.90 rows=1 width=12) (actual time=0.111..0.119 rows=1 loops=3209)
                                             Index Cond: (((groups_1."domain")::text = 'RT::Ticket-Role'::text) AND (("outer".id)::text = (groups_1.instance)::text) AND ((groups_1."type")::text = 'Requestor'::text))
                                 ->  Index Scan using principals2 on principals principals_2  (cost=0.00..4.62 rows=1 width=8) (actual time=0.015..0.018 rows=1 loops=3209)
                                       Index Cond: ("outer".id = principals_2.objectid)
                                       Filter: ((principaltype)::text = 'Group'::text)
                           ->  Index Scan using cachedgroupmembers3 on cachedgroupmembers cachedgroupmembers_3  (cost=0.00..134.06 rows=47 width=8) (actual time=0.015..0.026 rows=3 loops=3209)
                                 Index Cond: ("outer".id = cachedgroupmembers_3.groupid)
                     ->  Index Scan using users_pkey on users users_4  (cost=0.00..3.02 rows=1 width=4) (actual time=0.013..0.013 rows=0 loops=9617)
                           Index Cond: ("outer".memberid = users_4.id)
                           Filter: (lower((emailaddress)::text) = 'mallah_rajesh@yahoo.com'::text)
 Total runtime: 995.326 ms
(21 rows)
rt3=#
999 ms is not that bad but u think it deserves this many ms?


Nopes the query are not Equiv , earlier one returns 4 rows and the below one none,
can you spot any obvious and resend plz. thats why i did not do an explain analyze
rt3=# SELECT *
rt3-#   FROM tickets
rt3-#  WHERE id IN (
rt3(#        SELECT groups.instance
rt3(#          FROM groups
rt3(#          JOIN principals ON (groups.id = principals.objectid)
rt3(#          JOIN cachedgroupmembers ON (principals.id = cachedgroupmembers.groupid)
rt3(#          JOIN users ON (cachedgroupmembers.memberid = users.id)
rt3(#         WHERE users.emailaddress = 'mallah_rajesh@yahoo.com'
rt3(#           AND groups.domain = 'RT::Ticket-Role'
rt3(#           AND groups.type   = 'Requestor'
rt3(#           AND principals.principaltype = 'group'
rt3(#   )
rt3-#   AND type = 'ticket'
rt3-#   AND effectiveid = tickets.id
rt3-#   AND (status = 'new' OR status = 'open')
rt3-# ORDER BY priority DESC
rt3-# LIMIT 10;

 id | effectiveid | queue | type | issuestatement | resolution | owner | subject | initialpriority | finalpriority | priority | timeestimated | timeworked | status | timeleft | told | starts | started | due | resolved | lastupdatedby | lastupdated | creator | created | disabled
----+-------------+-------+------+----------------+------------+-------+---------+-----------------+---------------+----------+---------------+------------+--------+----------+------+--------+---------+-----+----------+---------------+-------------+---------+---------+----------
(0 rows)

Time: 2670.85 ms
rt3=#
Well it may be of interest to write the query in best possible way
but i am not sure if it really helps the RT application becoz i do
not know whether DBIx::SearchBuilder would currently allow
auto generation of such arbitrary SQLs.

Regds
Mallah.




Greg Stark wrote:
Rajesh Kumar Mallah <mallah@trade-india.com> writes:

  
rt3=# explain 

SELECT DISTINCT main.* 
  FROM (((
          (Tickets main  JOIN Groups as Groups_1 ON ( main.id = Groups_1.Instance))
          JOIN Principals as Principals_2 ON ( Groups_1.id = Principals_2.ObjectId)
         ) JOIN CachedGroupMembers as CachedGroupMembers_3  ON ( Principals_2.id = CachedGroupMembers_3.GroupId)
        ) JOIN Users as Users_4  ON ( CachedGroupMembers_3.MemberId = Users_4.id)
       )
 WHERE ((main.EffectiveId = main.id))
   AND ((main.Type = 'ticket'))
   AND (((    (Users_4.EmailAddress = 'mallah_rajesh@yahoo.com')
          AND (Groups_1.Domain = 'RT::Ticket-Role')
          AND (Groups_1.Type = 'Requestor')
          AND (Principals_2.PrincipalType = 'Group')
         ))
         AND ((main.Status = 'new') OR (main.Status = 'open'))
       )
 ORDER BY main.Priority DESC LIMIT 10;
    

So this query seems to be going the long way around to do the equivalent of an
IN clause. Presumably because as far as I know mysql didn't support IN
subqueries until recently.

Can you do an "explain analyze" on the above query and the following rewritten
one in 7.4? The "analyze" is important because it'll give real timing
information. And it's important that it be on 7.4 as there were improvements
in this area specifically in 7.4.

SELECT * 
  FROM tickets
 WHERE id IN (
       SELECT groups.instance
         FROM groups 
         JOIN principals ON (groups.id = principals.objectid) 
         JOIN cachedgroupmembers ON (principals.id = cachedgroupmembers.groupid)
         JOIN users ON (cachedgroupmembers.memberid = users.id)
        WHERE users.emailaddress = 'mallah_rajesh@yahoo.com'
          AND groups.domain = 'RT::Ticket-Role'
          AND groups.type   = 'Requestor'
          AND principals.principaltype = 'group'
  )
  AND type = 'ticket'
  AND effectiveid = tickets.id 
  AND (status = 'new' OR status = 'open')
ORDER BY priority DESC 
LIMIT 10;

--------------070306070203080407090902-- From pgsql-performance-owner@postgresql.org Sat Nov 1 01:48:23 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A9938D1B53E for ; Sat, 1 Nov 2003 05:48:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 57325-08 for ; Sat, 1 Nov 2003 01:47:51 -0400 (AST) Received: from rtlocal.trade-india.com (unknown [61.16.154.82]) by svr1.postgresql.org (Postfix) with SMTP id A6EA1D1B509 for ; Sat, 1 Nov 2003 01:47:47 -0400 (AST) Received: (qmail 5547 invoked from network); 1 Nov 2003 06:04:40 -0000 Received: from unknown (HELO trade-india.com) (unknown) by unknown with SMTP; 1 Nov 2003 06:04:40 -0000 From: Rajesh Kumar Mallah Organization: Infocom Network Limited To: Greg Stark Subject: [ PROBLEM SOLVED ] Re: Query puts 7.3.4 on endless loop but 7.4beta5 is fine. Date: Sat, 1 Nov 2003 11:17:02 +0530 User-Agent: KMail/1.5.1 Cc: Christopher Browne , pgsql-performance@postgresql.org, Jesse Vincent References: <33298.203.145.130.142.1067457626.squirrel@mail.trade-india.com> <3FA34827.1030705@trade-india.com> <87vfq6bbbn.fsf@stark.dyndns.tv> In-Reply-To: <87vfq6bbbn.fsf@stark.dyndns.tv> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311011117.02662.mallah@trade-india.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/1 X-Sequence-Number: 4561 Well Sorry everyone , The problem was tracked down to a silly datatype mismatch between two join columns in table Groups(instance) and Tickets(id) (int vs varchar ) 7.4b5 is automatically taking care of this mismatch hence it was getting executed there. But , The problem is will this behaviour not allow to go such mistakes unnoticed? Regards Mallah. On Friday 31 Oct 2003 4:08 am, Greg Stark wrote: > Well, you might want to try the EXISTS version. I'm not sure if it'll be > faster or slower though. In theory it should be the same. > > Hum, I didn't realize the principals table was the largest table. But > Postgres knew that so one would expect it to have found a better plan. The > IN/EXISTS handling was recently much improved but perhaps there's still > room :) > > SELECT * > FROM tickets > WHERE EXISTS ( > SELECT 1 > FROM groups > JOIN principals ON (groups.id = principals.objectid) > JOIN cachedgroupmembers ON (principals.id = > cachedgroupmembers.groupid) JOIN users ON (cachedgroupmembers.memberid = > users.id) > WHERE lower(users.emailaddress) = 'mallah_rajesh@yahoo.com' > AND groups.domain = 'RT::Ticket-Role' > AND groups.type = 'Requestor' > AND principals.principaltype = 'group' > AND groups.instance = tickets.id > ) > AND type = 'ticket' > AND effectiveid = tickets.id > AND (status = 'new' OR status = 'open') > ORDER BY priority DESC > LIMIT 10; From pgsql-performance-owner@postgresql.org Thu Oct 30 18:01:43 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 910DAD1B4E9 for ; Thu, 30 Oct 2003 22:01:39 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 44529-02 for ; Thu, 30 Oct 2003 18:01:09 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id A5467D1B511 for ; Thu, 30 Oct 2003 18:01:06 -0400 (AST) Received: (qmail 25831 invoked from network); 30 Oct 2003 22:02:26 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 30 Oct 2003 22:02:26 -0000 Message-ID: <3FA36F4C.2000509@trade-india.com> Date: Sat, 01 Nov 2003 03:31:08 -0500 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Greg Stark Cc: Christopher Browne , pgsql-performance@postgresql.org, Jesse Vincent Subject: Re: Query puts 7.3.4 on endless loop but 7.4beta5 is fine. References: <33298.203.145.130.142.1067457626.squirrel@mail.trade-india.com> <20031029202920.GZ7337@pallas.fsck.com> <33402.203.145.130.142.1067460429.squirrel@mail.trade-india.com> <200310301702.00275.mallah@trade-india.com> <13042.61.16.154.82.1067523120.squirrel@mail.trade-india.com> <3FA323C2.2060302@trade-india.com> <87r80ud0nw.fsf@stark.dyndns.tv> <3FA34827.1030705@trade-india.com> <87ad7icsxw.fsf@stark.dyndns.tv> In-Reply-To: <87ad7icsxw.fsf@stark.dyndns.tv> Content-Type: multipart/alternative; boundary="------------040101020809020906010105" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200310/780 X-Sequence-Number: 4528 This is a multi-part message in MIME format. --------------040101020809020906010105 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit The g in group had to be uppercased, the query produced the same results but performance was worse for the IN version . 2367 ms vs 600 ms rt3=# explain analyze SELECT * from tickets where id in ( SELECT groups.instance FROM groups JOIN principals ON (groups.id = principals.objectid) JOIN cachedgroupmembers ON (principals.id = cachedgroupmembers.groupid) JOIN users ON (cachedgroupmembers.memberid = users.id) WHERE lower(users.emailaddress) = 'mallah_rajesh@yahoo.com' AND groups.domain = 'RT::Ticket-Role' AND groups.type = 'Requestor' AND principals.principaltype = 'Group' ) AND type = 'ticket' AND effectiveid = tickets.id AND (status = 'new' OR status = 'open') ORDER BY priority DESC LIMIT 10;; QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Limit (cost=10078.18..10078.19 rows=1 width=164) (actual time=2367.084..2367.096 rows=4 loops=1) -> Sort (cost=10078.18..10078.19 rows=1 width=164) (actual time=2367.078..2367.082 rows=4 loops=1) Sort Key: tickets.priority -> Hash Join (cost=10077.65..10078.17 rows=1 width=164) (actual time=2366.870..2367.051 rows=4 loops=1) Hash Cond: (("outer".instance)::text = ("inner".id)::text) -> HashAggregate (cost=9612.02..9612.02 rows=69 width=8) (actual time=2303.792..2303.810 rows=7 loops=1) -> Hash Join (cost=4892.97..9611.85 rows=69 width=8) (actual time=1427.260..2303.685 rows=14 loops=1) Hash Cond: ("outer".memberid = "inner".id) -> Hash Join (cost=4523.65..9139.45 rows=13651 width=12) (actual time=948.960..2258.529 rows=31123 loops=1) Hash Cond: ("outer".groupid = "inner".id) -> Seq Scan on cachedgroupmembers (cost=0.00..3456.51 rows=204551 width=8) (actual time=0.048..365.147 rows=204551 loops=1) -> Hash (cost=4509.93..4509.93 rows=5488 width=12) (actual time=948.843..948.843 rows=0 loops=1) -> Hash Join (cost=1409.91..4509.93 rows=5488 width=12) (actual time=315.722..930.025 rows=10431 loops=1) Hash Cond: ("outer".objectid = "inner".id) -> Seq Scan on principals (cost=0.00..1583.76 rows=62625 width=8) (actual time=0.043..251.142 rows=62097 loops=1) Filter: ((principaltype)::text = 'Group'::text) -> Hash (cost=1359.90..1359.90 rows=7204 width=12) (actual time=315.458..315.458 rows=0 loops=1) -> Index Scan using groups_domain on groups (cost=0.00..1359.90 rows=7204 width=12) (actual time=0.325..297.403 rows=10431 loops=1) Index Cond: (("domain")::text = 'RT::Ticket-Role'::text) Filter: (("type")::text = 'Requestor'::text) -> Hash (cost=369.08..369.08 rows=101 width=4) (actual time=0.157..0.157 rows=0 loops=1) -> Index Scan using users_emailaddress_lower on users (cost=0.00..369.08 rows=101 width=4) (actual time=0.139..0.143 rows=1 loops=1) Index Cond: (lower((emailaddress)::text) = 'mallah_rajesh@yahoo.com'::text) -> Hash (cost=465.62..465.62 rows=1 width=164) (actual time=62.944..62.944 rows=0 loops=1) -> Seq Scan on tickets (cost=0.00..465.62 rows=1 width=164) (actual time=0.113..52.729 rows=3208 loops=1) Filter: ((("type")::text = 'ticket'::text) AND (effectiveid = id) AND (((status)::text = 'new'::text) OR ((status)::text = 'open'::text))) Total runtime: 2367.908 ms (27 rows) rt3=# explain analyze SELECT DISTINCT main.* FROM ((((Tickets main JOIN Groups as Groups_1 ON ( main.id = Groups_1.Instance)) rt3(# JOIN Principals as Principals_2 ON ( Groups_1.id = Principals_2.ObjectId)) JOIN CachedGroupMembers as CachedGroupMembers_3 rt3(# ON ( Principals_2.id = CachedGroupMembers_3.GroupId)) JOIN Users as Users_4 ON ( CachedGroupMembers_3.MemberId = Users_4.id)) rt3-# WHERE ((main.EffectiveId = main.id)) AND ((main.Type = 'ticket')) AND ( ( ( (Users_4.EmailAddress = 'mallah_rajesh@yahoo.com') rt3(# AND(Groups_1.Domain = 'RT::Ticket-Role')AND(Groups_1.Type = 'Requestor')AND(Principals_2.PrincipalType = 'Group') ) ) rt3(# AND ( (main.Status = 'new')OR(main.Status = 'open') ) ) ORDER BY main.Priority DESC LIMIT 10; QUERY PLAN ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ Limit (cost=582.27..582.34 rows=1 width=164) (actual time=592.406..592.529 rows=4 loops=1) -> Unique (cost=582.27..582.34 rows=1 width=164) (actual time=592.401..592.516 rows=4 loops=1) -> Sort (cost=582.27..582.28 rows=1 width=164) (actual time=592.398..592.406 rows=8 loops=1) Sort Key: main.priority, main.id, main.effectiveid, main.queue, main."type", main.issuestatement, main.resolution, main."owner", main.subject, main.initialpriority, main.finalpriority, main.timeestimated, main.timeworked, main.status, main.timeleft, main.told, main.starts, main.started, main.due, main.resolved, main.lastupdatedby, main.lastupdated, main.creator, main.created, main.disabled -> Hash Join (cost=476.18..582.26 rows=1 width=164) (actual time=591.548..592.211 rows=8 loops=1) Hash Cond: ("outer".groupid = "inner".id) -> Nested Loop (cost=0.00..105.97 rows=21 width=4) (actual time=0.214..0.645 rows=37 loops=1) -> Index Scan using users4 on users users_4 (cost=0.00..3.99 rows=2 width=4) (actual time=0.107..0.112 rows=1 loops=1) Index Cond: ((emailaddress)::text = 'mallah_rajesh@yahoo.com'::text) -> Index Scan using cachedgroupmembers2 on cachedgroupmembers cachedgroupmembers_3 (cost=0.00..50.81 rows=14 width=8) (actual time=0.098..0.441 rows=37 loops=1) Index Cond: (cachedgroupmembers_3.memberid = "outer".id) -> Hash (cost=476.17..476.17 rows=1 width=168) (actual time=591.121..591.121 rows=0 loops=1) -> Nested Loop (cost=0.00..476.17 rows=1 width=168) (actual time=0.391..583.085 rows=3208 loops=1) -> Nested Loop (cost=0.00..471.54 rows=1 width=168) (actual time=0.309..474.968 rows=3208 loops=1) -> Seq Scan on tickets main (cost=0.00..465.62 rows=1 width=164) (actual time=0.111..56.930 rows=3208 loops=1) Filter: ((effectiveid = id) AND (("type")::text = 'ticket'::text) AND (((status)::text = 'new'::text) OR ((status)::text = 'open'::text))) -> Index Scan using groups1 on groups groups_1 (cost=0.00..5.90 rows=1 width=12) (actual time=0.105..0.112 rows=1 loops=3208) Index Cond: (((groups_1."domain")::text = 'RT::Ticket-Role'::text) AND (("outer".id)::text = (groups_1.instance)::text) AND ((groups_1."type")::text = 'Requestor'::text)) -> Index Scan using principals2 on principals principals_2 (cost=0.00..4.62 rows=1 width=8) (actual time=0.014..0.017 rows=1 loops=3208) Index Cond: ("outer".id = principals_2.objectid) Filter: ((principaltype)::text = 'Group'::text) Total runtime: 593.062 ms (22 rows) Regds Mallah. Greg Stark wrote: >Rajesh Kumar Mallah writes: > > > >>Nopes the query are not Equiv , earlier one returns 4 rows and the below one >>none, >> >> > >Sorry, i lowercased a string constant and dropped the lower() on email. > >Try this: > >SELECT * > FROM tickets > WHERE id IN ( > SELECT groups.instance > FROM groups > JOIN principals ON (groups.id = principals.objectid) > JOIN cachedgroupmembers ON (principals.id = cachedgroupmembers.groupid) > JOIN users ON (cachedgroupmembers.memberid = users.id) > WHERE lower(users.emailaddress) = 'mallah_rajesh@yahoo.com' > AND groups.domain = 'RT::Ticket-Role' > AND groups.type = 'Requestor' > AND principals.principaltype = 'group' > ) > AND type = 'ticket' > AND effectiveid = tickets.id > AND (status = 'new' OR status = 'open') >ORDER BY priority DESC >LIMIT 10; > > > --------------040101020809020906010105 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit
The g in group had to be uppercased, the query produced the same results
but performance was worse  for the IN version .  2367 ms vs 600 ms
rt3=# explain analyze SELECT  * from tickets where id in (  SELECT groups.instance FROM groups
 JOIN principals ON (groups.id = principals.objectid) JOIN cachedgroupmembers ON 
(principals.id = cachedgroupmembers.groupid) JOIN users ON (cachedgroupmembers.memberid = users.id)  
WHERE lower(users.emailaddress) = 'mallah_rajesh@yahoo.com' AND groups.domain = 'RT::Ticket-Role' 
AND groups.type   = 'Requestor' AND principals.principaltype = 'Group' ) AND type = 'ticket' AND 
effectiveid = tickets.id AND (status = 'new' OR status = 'open') ORDER BY priority DESC LIMIT 10;;

                                                                                       QUERY PLAN
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 Limit  (cost=10078.18..10078.19 rows=1 width=164) (actual time=2367.084..2367.096 rows=4 loops=1)
   ->  Sort  (cost=10078.18..10078.19 rows=1 width=164) (actual time=2367.078..2367.082 rows=4 loops=1)
         Sort Key: tickets.priority
         ->  Hash Join  (cost=10077.65..10078.17 rows=1 width=164) (actual time=2366.870..2367.051 rows=4 loops=1)
               Hash Cond: (("outer".instance)::text = ("inner".id)::text)
               ->  HashAggregate  (cost=9612.02..9612.02 rows=69 width=8) (actual time=2303.792..2303.810 rows=7 loops=1)
                     ->  Hash Join  (cost=4892.97..9611.85 rows=69 width=8) (actual time=1427.260..2303.685 rows=14 loops=1)
                           Hash Cond: ("outer".memberid = "inner".id)
                           ->  Hash Join  (cost=4523.65..9139.45 rows=13651 width=12) (actual time=948.960..2258.529 rows=31123 loops=1)
                                 Hash Cond: ("outer".groupid = "inner".id)
                                 ->  Seq Scan on cachedgroupmembers  (cost=0.00..3456.51 rows=204551 width=8) (actual time=0.048..365.147 rows=204551 loops=1)
                                 ->  Hash  (cost=4509.93..4509.93 rows=5488 width=12) (actual time=948.843..948.843 rows=0 loops=1)
                                       ->  Hash Join  (cost=1409.91..4509.93 rows=5488 width=12) (actual time=315.722..930.025 rows=10431 loops=1)
                                             Hash Cond: ("outer".objectid = "inner".id)
                                             ->  Seq Scan on principals  (cost=0.00..1583.76 rows=62625 width=8) (actual time=0.043..251.142 rows=62097 loops=1)
                                                   Filter: ((principaltype)::text = 'Group'::text)
                                             ->  Hash  (cost=1359.90..1359.90 rows=7204 width=12) (actual time=315.458..315.458 rows=0 loops=1)
                                                   ->  Index Scan using groups_domain on groups  (cost=0.00..1359.90 rows=7204 width=12) (actual time=0.325..297.403 rows=10431 loops=1)
                                                         Index Cond: (("domain")::text = 'RT::Ticket-Role'::text)
                                                         Filter: (("type")::text = 'Requestor'::text)
                           ->  Hash  (cost=369.08..369.08 rows=101 width=4) (actual time=0.157..0.157 rows=0 loops=1)
                                 ->  Index Scan using users_emailaddress_lower on users  (cost=0.00..369.08 rows=101 width=4) (actual time=0.139..0.143 rows=1 loops=1)
                                       Index Cond: (lower((emailaddress)::text) = 'mallah_rajesh@yahoo.com'::text)
               ->  Hash  (cost=465.62..465.62 rows=1 width=164) (actual time=62.944..62.944 rows=0 loops=1)
                     ->  Seq Scan on tickets  (cost=0.00..465.62 rows=1 width=164) (actual time=0.113..52.729 rows=3208 loops=1)
                           Filter: ((("type")::text = 'ticket'::text) AND (effectiveid = id) AND (((status)::text = 'new'::text) OR ((status)::text = 'open'::text)))
 Total runtime: 2367.908 ms
(27 rows)



rt3=# explain analyze SELECT DISTINCT main.* FROM ((((Tickets main  JOIN Groups as Groups_1  ON ( main.id = Groups_1.Instance))
rt3(# JOIN Principals as Principals_2  ON ( Groups_1.id = Principals_2.ObjectId)) JOIN CachedGroupMembers as CachedGroupMembers_3
rt3(# ON ( Principals_2.id = CachedGroupMembers_3.GroupId))  JOIN Users as Users_4  ON ( CachedGroupMembers_3.MemberId = Users_4.id))
rt3-# WHERE ((main.EffectiveId = main.id)) AND ((main.Type = 'ticket')) AND ( (  ( (Users_4.EmailAddress = 'mallah_rajesh@yahoo.com')
rt3(# AND(Groups_1.Domain = 'RT::Ticket-Role')AND(Groups_1.Type = 'Requestor')AND(Principals_2.PrincipalType = 'Group') )  )
rt3(# AND ( (main.Status = 'new')OR(main.Status = 'open') ) )  ORDER BY main.Priority DESC LIMIT 10;


                                                                                                                                                                                                       QUERY PLAN
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 Limit  (cost=582.27..582.34 rows=1 width=164) (actual time=592.406..592.529 rows=4 loops=1)
   ->  Unique  (cost=582.27..582.34 rows=1 width=164) (actual time=592.401..592.516 rows=4 loops=1)
         ->  Sort  (cost=582.27..582.28 rows=1 width=164) (actual time=592.398..592.406 rows=8 loops=1)
               Sort Key: main.priority, main.id, main.effectiveid, main.queue, main."type", main.issuestatement, main.resolution, main."owner", main.subject, main.initialpriority, main.finalpriority, main.timeestimated, main.timeworked, main.status, main.timeleft, main.told, main.starts, main.started, main.due, main.resolved, main.lastupdatedby, main.lastupdated, main.creator, main.created, main.disabled
               ->  Hash Join  (cost=476.18..582.26 rows=1 width=164) (actual time=591.548..592.211 rows=8 loops=1)
                     Hash Cond: ("outer".groupid = "inner".id)
                     ->  Nested Loop  (cost=0.00..105.97 rows=21 width=4) (actual time=0.214..0.645 rows=37 loops=1)
                           ->  Index Scan using users4 on users users_4  (cost=0.00..3.99 rows=2 width=4) (actual time=0.107..0.112 rows=1 loops=1)
                                 Index Cond: ((emailaddress)::text = 'mallah_rajesh@yahoo.com'::text)
                           ->  Index Scan using cachedgroupmembers2 on cachedgroupmembers cachedgroupmembers_3  (cost=0.00..50.81 rows=14 width=8) (actual time=0.098..0.441 rows=37 loops=1)
                                 Index Cond: (cachedgroupmembers_3.memberid = "outer".id)
                     ->  Hash  (cost=476.17..476.17 rows=1 width=168) (actual time=591.121..591.121 rows=0 loops=1)
                           ->  Nested Loop  (cost=0.00..476.17 rows=1 width=168) (actual time=0.391..583.085 rows=3208 loops=1)
                                 ->  Nested Loop  (cost=0.00..471.54 rows=1 width=168) (actual time=0.309..474.968 rows=3208 loops=1)
                                       ->  Seq Scan on tickets main  (cost=0.00..465.62 rows=1 width=164) (actual time=0.111..56.930 rows=3208 loops=1)
                                             Filter: ((effectiveid = id) AND (("type")::text = 'ticket'::text) AND (((status)::text = 'new'::text) OR ((status)::text = 'open'::text)))
                                       ->  Index Scan using groups1 on groups groups_1  (cost=0.00..5.90 rows=1 width=12) (actual time=0.105..0.112 rows=1 loops=3208)
                                             Index Cond: (((groups_1."domain")::text = 'RT::Ticket-Role'::text) AND (("outer".id)::text = (groups_1.instance)::text) AND ((groups_1."type")::text = 'Requestor'::text))
                                 ->  Index Scan using principals2 on principals principals_2  (cost=0.00..4.62 rows=1 width=8) (actual time=0.014..0.017 rows=1 loops=3208)
                                       Index Cond: ("outer".id = principals_2.objectid)
                                       Filter: ((principaltype)::text = 'Group'::text)
 Total runtime: 593.062 ms
(22 rows)
Regds
Mallah.

Greg Stark wrote:
Rajesh Kumar Mallah <mallah@trade-india.com> writes:

  
Nopes the query are not Equiv , earlier one returns 4 rows and the below one
none,
    

Sorry, i lowercased a string constant and dropped the lower() on email. 

Try this:

SELECT *
  FROM tickets
 WHERE id IN (
       SELECT groups.instance
         FROM groups
         JOIN principals ON (groups.id = principals.objectid)
         JOIN cachedgroupmembers ON (principals.id = cachedgroupmembers.groupid)
         JOIN users ON (cachedgroupmembers.memberid = users.id)
        WHERE lower(users.emailaddress) = 'mallah_rajesh@yahoo.com'
          AND groups.domain = 'RT::Ticket-Role'
          AND groups.type   = 'Requestor'
          AND principals.principaltype = 'group'
  )
  AND type = 'ticket'
  AND effectiveid = tickets.id
  AND (status = 'new' OR status = 'open')
ORDER BY priority DESC
LIMIT 10;

--------------040101020809020906010105-- From pgsql-performance-owner@postgresql.org Thu Oct 30 18:21:32 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B13F1D1B50E for ; Thu, 30 Oct 2003 22:21:30 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 44626-10 for ; Thu, 30 Oct 2003 18:21:00 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id C57A2D1B8A0 for ; Thu, 30 Oct 2003 18:20:57 -0400 (AST) Received: (qmail 27522 invoked from network); 30 Oct 2003 22:22:17 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 30 Oct 2003 22:22:17 -0000 Message-ID: <3FA373F3.9040103@trade-india.com> Date: Sat, 01 Nov 2003 03:50:59 -0500 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Greg Stark Cc: Christopher Browne , pgsql-performance@postgresql.org, Jesse Vincent Subject: Re: Query puts 7.3.4 on endless loop but 7.4beta5 is fine. References: <33298.203.145.130.142.1067457626.squirrel@mail.trade-india.com> <20031029202920.GZ7337@pallas.fsck.com> <33402.203.145.130.142.1067460429.squirrel@mail.trade-india.com> <200310301702.00275.mallah@trade-india.com> <13042.61.16.154.82.1067523120.squirrel@mail.trade-india.com> <3FA323C2.2060302@trade-india.com> <87r80ud0nw.fsf@stark.dyndns.tv> <3FA34827.1030705@trade-india.com> <87ad7icsxw.fsf@stark.dyndns.tv> In-Reply-To: <87ad7icsxw.fsf@stark.dyndns.tv> Content-Type: multipart/alternative; boundary="------------060402080107040008000108" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200310/781 X-Sequence-Number: 4529 This is a multi-part message in MIME format. --------------060402080107040008000108 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit But the new version at lease works on 7.3 instead of putting it in an infinite loop. rt3=# explain analyze SELECT * from tickets where id in ( SELECT groups.instance FROM groups rt3(# JOIN principals ON (groups.id = principals.objectid) JOIN cachedgroupmembers ON rt3(# (principals.id = cachedgroupmembers.groupid) JOIN users ON (cachedgroupmembers.memberid = users.id) rt3(# WHERE lower(users.emailaddress) = 'mallah_rajesh@yahoo.com' AND groups.domain = 'RT::Ticket-Role' rt3(# AND groups.type = 'Requestor' AND principals.principaltype = 'Group' ) AND type = 'ticket' AND rt3-# effectiveid = tickets.id AND (status = 'new' OR status = 'open') ORDER BY priority DESC LIMIT 10; QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Limit (cost=88073404.73..88073404.73 rows=1 width=163) (actual time=2859.05..2859.07 rows=4 loops=1) -> Sort (cost=88073404.73..88073404.73 rows=1 width=163) (actual time=2859.05..2859.05 rows=4 loops=1) Sort Key: priority -> Seq Scan on tickets (cost=0.00..88073404.72 rows=1 width=163) (actual time=2525.48..2858.95 rows=4 loops=1) Filter: (("type" = 'ticket'::character varying) AND (effectiveid = id) AND ((status = 'new'::character varying) OR (status = 'open'::character varying)) AND (subplan)) SubPlan -> Materialize (cost=8443.38..8443.38 rows=66 width=32) (actual time=0.79..0.81 rows=14 loops=3209) -> Hash Join (cost=3698.35..8443.38 rows=66 width=32) (actual time=1720.53..2525.07 rows=14 loops=1) Hash Cond: ("outer".memberid = "inner".id) -> Hash Join (cost=3329.03..7973.87 rows=13247 width=28) (actual time=1225.83..2458.48 rows=31123 loops=1) Hash Cond: ("outer".groupid = "inner".id) -> Seq Scan on cachedgroupmembers (cost=0.00..3456.51 rows=204551 width=8) (actual time=0.06..638.91 rows=204551 loops=1) -> Hash (cost=3315.71..3315.71 rows=5325 width=20) (actual time=1225.51..1225.51 rows=0 loops=1) -> Hash Join (cost=1355.70..3315.71 rows=5325 width=20) (actual time=529.02..1191.94 rows=10431 loops=1) Hash Cond: ("outer".objectid = "inner".id) -> Seq Scan on principals (cost=0.00..1583.76 rows=61940 width=8) (actual time=0.02..450.42 rows=62097 loops=1) Filter: (principaltype = 'Group'::character varying) -> Hash (cost=1338.03..1338.03 rows=7068 width=12) (actual time=528.58..528.58 rows=0 loops=1) -> Index Scan using groups_domain on groups (cost=0.00..1338.03 rows=7068 width=12) (actual time=0.18..498.04 rows=10431 loops=1) Index Cond: ("domain" = 'RT::Ticket-Role'::character varying) Filter: ("type" = 'Requestor'::character varying) -> Hash (cost=369.08..369.08 rows=101 width=4) (actual time=0.10..0.10 rows=0 loops=1) -> Index Scan using users_emailaddress on users (cost=0.00..369.08 rows=101 width=4) (actual time=0.09..0.10 rows=1 loops=1) Index Cond: (lower((emailaddress)::text) = 'mallah_rajesh@yahoo.com'::text) Total runtime: 2859.34 msec (25 rows) Greg Stark wrote: >Rajesh Kumar Mallah writes: > > > >>Nopes the query are not Equiv , earlier one returns 4 rows and the below one >>none, >> >> > >Sorry, i lowercased a string constant and dropped the lower() on email. > >Try this: > >SELECT * > FROM tickets > WHERE id IN ( > SELECT groups.instance > FROM groups > JOIN principals ON (groups.id = principals.objectid) > JOIN cachedgroupmembers ON (principals.id = cachedgroupmembers.groupid) > JOIN users ON (cachedgroupmembers.memberid = users.id) > WHERE lower(users.emailaddress) = 'mallah_rajesh@yahoo.com' > AND groups.domain = 'RT::Ticket-Role' > AND groups.type = 'Requestor' > AND principals.principaltype = 'group' > ) > AND type = 'ticket' > AND effectiveid = tickets.id > AND (status = 'new' OR status = 'open') >ORDER BY priority DESC >LIMIT 10; > > > --------------060402080107040008000108 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit
But the new version at lease works on 7.3 instead of putting
it in an infinite loop.


rt3=# explain analyze SELECT  * from tickets where id in (  SELECT groups.instance FROM groups
rt3(#  JOIN principals ON (groups.id = principals.objectid) JOIN cachedgroupmembers ON
rt3(# (principals.id = cachedgroupmembers.groupid) JOIN users ON (cachedgroupmembers.memberid = users.id)
rt3(# WHERE lower(users.emailaddress) = 'mallah_rajesh@yahoo.com' AND groups.domain = 'RT::Ticket-Role'
rt3(# AND groups.type   = 'Requestor' AND principals.principaltype = 'Group' ) AND type = 'ticket' AND
rt3-# effectiveid = tickets.id AND (status = 'new' OR status = 'open') ORDER BY priority DESC LIMIT 10;


                                                                                       QUERY PLAN
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
 Limit  (cost=88073404.73..88073404.73 rows=1 width=163) (actual time=2859.05..2859.07 rows=4 loops=1)
   ->  Sort  (cost=88073404.73..88073404.73 rows=1 width=163) (actual time=2859.05..2859.05 rows=4 loops=1)
         Sort Key: priority
         ->  Seq Scan on tickets  (cost=0.00..88073404.72 rows=1 width=163) (actual time=2525.48..2858.95 rows=4 loops=1)
               Filter: (("type" = 'ticket'::character varying) AND (effectiveid = id) AND ((status = 'new'::character varying) OR (status = 'open'::character varying)) AND (subplan))
               SubPlan
                 ->  Materialize  (cost=8443.38..8443.38 rows=66 width=32) (actual time=0.79..0.81 rows=14 loops=3209)
                       ->  Hash Join  (cost=3698.35..8443.38 rows=66 width=32) (actual time=1720.53..2525.07 rows=14 loops=1)
                             Hash Cond: ("outer".memberid = "inner".id)
                             ->  Hash Join  (cost=3329.03..7973.87 rows=13247 width=28) (actual time=1225.83..2458.48 rows=31123 loops=1)
                                   Hash Cond: ("outer".groupid = "inner".id)
                                   ->  Seq Scan on cachedgroupmembers  (cost=0.00..3456.51 rows=204551 width=8) (actual time=0.06..638.91 rows=204551 loops=1)
                                   ->  Hash  (cost=3315.71..3315.71 rows=5325 width=20) (actual time=1225.51..1225.51 rows=0 loops=1)
                                         ->  Hash Join  (cost=1355.70..3315.71 rows=5325 width=20) (actual time=529.02..1191.94 rows=10431 loops=1)
                                               Hash Cond: ("outer".objectid = "inner".id)
                                               ->  Seq Scan on principals  (cost=0.00..1583.76 rows=61940 width=8) (actual time=0.02..450.42 rows=62097 loops=1)
                                                     Filter: (principaltype = 'Group'::character varying)
                                               ->  Hash  (cost=1338.03..1338.03 rows=7068 width=12) (actual time=528.58..528.58 rows=0 loops=1)
                                                     ->  Index Scan using groups_domain on groups  (cost=0.00..1338.03 rows=7068 width=12) (actual time=0.18..498.04 rows=10431 loops=1)
                                                           Index Cond: ("domain" = 'RT::Ticket-Role'::character varying)
                                                           Filter: ("type" = 'Requestor'::character varying)
                             ->  Hash  (cost=369.08..369.08 rows=101 width=4) (actual time=0.10..0.10 rows=0 loops=1)
                                   ->  Index Scan using users_emailaddress on users  (cost=0.00..369.08 rows=101 width=4) (actual time=0.09..0.10 rows=1 loops=1)
                                         Index Cond: (lower((emailaddress)::text) = 'mallah_rajesh@yahoo.com'::text)
 Total runtime: 2859.34 msec
(25 rows)





Greg Stark wrote:
Rajesh Kumar Mallah <mallah@trade-india.com> writes:

  
Nopes the query are not Equiv , earlier one returns 4 rows and the below one
none,
    

Sorry, i lowercased a string constant and dropped the lower() on email. 

Try this:

SELECT *
  FROM tickets
 WHERE id IN (
       SELECT groups.instance
         FROM groups
         JOIN principals ON (groups.id = principals.objectid)
         JOIN cachedgroupmembers ON (principals.id = cachedgroupmembers.groupid)
         JOIN users ON (cachedgroupmembers.memberid = users.id)
        WHERE lower(users.emailaddress) = 'mallah_rajesh@yahoo.com'
          AND groups.domain = 'RT::Ticket-Role'
          AND groups.type   = 'Requestor'
          AND principals.principaltype = 'group'
  )
  AND type = 'ticket'
  AND effectiveid = tickets.id
  AND (status = 'new' OR status = 'open')
ORDER BY priority DESC
LIMIT 10;

--------------060402080107040008000108-- From pgsql-performance-owner@postgresql.org Sun Nov 2 12:33:45 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 37424D1B50B for ; Sun, 2 Nov 2003 16:33:42 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 64769-05 for ; Sun, 2 Nov 2003 12:33:10 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 4D4D7D1B554 for ; Sun, 2 Nov 2003 12:33:10 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id 3AC71369B6; Sun, 2 Nov 2003 11:33:04 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1AGLA0-0008WA-00; Sun, 02 Nov 2003 11:33:04 -0500 To: William Yu Cc: pgsql-performance@postgresql.org Subject: Re: Pg+Linux swap use References: <20031031143624.GB5394@libertyrms.info> <007c01c39fc4$0844f850$640101c0@rob> In-Reply-To: From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 02 Nov 2003 11:33:03 -0500 Message-ID: <87y8uy91ds.fsf@stark.dyndns.tv> Lines: 22 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/4 X-Sequence-Number: 4564 William Yu writes: > Rob Sell wrote: > > > Not being one to hijack threads, but I haven't heard of this performance hit > > when using HT, I have what should all rights be a pretty fast server, dual > > 2.4 Xeons with HT 205gb raid 5 array, 1 gig of memory. And it is only 50% as > > fast as my old server which was a dual AMD MP 1400's with a 45gb raid 5 > > array and 1gb of ram. > > Not to get into a big Intel vs AMD argument but 50% sounds about right. Let's > first assume that the QS rating for the MP1400 is relatively accurate and > convert that to a 1.4GHz Xeon. 2.4/1.4 = +71%. Since processor performance > does not increase linearly with clockspeed, 50% is in line with expectations. Hm. You've read "50% as fast" as "50% faster". I wonder which the original poster intended. -- greg From pgsql-performance-owner@postgresql.org Mon Nov 3 12:17:22 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 04C98D1B58D for ; Mon, 3 Nov 2003 16:17:20 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 86820-07 for ; Mon, 3 Nov 2003 12:16:49 -0400 (AST) Received: from mail.owensforestproducts.com (unknown [216.161.134.233]) by svr1.postgresql.org (Postfix) with ESMTP id A2175D1B52F for ; Mon, 3 Nov 2003 12:16:48 -0400 (AST) Subject: join_collapse_limit, from_collapse_limit options missing MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Date: Mon, 3 Nov 2003 10:20:53 -0600 content-class: urn:content-classes:message X-MimeOLE: Produced By Microsoft Exchange V6.0.6249.0 Message-ID: <70AA9BD4C82E76418F730DF1411A624F06DF12@meta01.owens.int> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: join_collapse_limit, from_collapse_limit options missing thread-index: AcOiJnPKfSaPwj1oT8OZmc1hdBt5VA== From: "Lee Hughes" To: X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/5 X-Sequence-Number: 4565 Hi, I'm trying to set these two options to tune performance but both return 'not a valid option name'. Dumping the pg_settings table confirms that they are missing. I'm using the PostgreSQL packages included with RedHat 9 (7.3.2) and Mandrake 9.2 Beta (7.3.4). Do I need to install the full tarball to get these options? Any help greatly appreciated. Thanks, Lee From pgsql-performance-owner@postgresql.org Mon Nov 3 12:25:07 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 52C46D1CAC1 for ; Mon, 3 Nov 2003 16:25:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 95254-01 for ; Mon, 3 Nov 2003 12:24:31 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id A1739D1CAB7 for ; Mon, 3 Nov 2003 12:24:30 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hA3GOP19028122; Mon, 3 Nov 2003 11:24:26 -0500 (EST) To: "Lee Hughes" Cc: pgsql-performance@postgresql.org Subject: Re: join_collapse_limit, from_collapse_limit options missing In-reply-to: <70AA9BD4C82E76418F730DF1411A624F06DF12@meta01.owens.int> References: <70AA9BD4C82E76418F730DF1411A624F06DF12@meta01.owens.int> Comments: In-reply-to "Lee Hughes" message dated "Mon, 03 Nov 2003 10:20:53 -0600" Date: Mon, 03 Nov 2003 11:24:25 -0500 Message-ID: <28121.1067876665@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/6 X-Sequence-Number: 4566 "Lee Hughes" writes: > Hi, I'm trying to set these two options to tune performance but both > return 'not a valid option name'. Dumping the pg_settings table confirms > that they are missing. I'm using the PostgreSQL packages included with > RedHat 9 (7.3.2) and Mandrake 9.2 Beta (7.3.4). Do I need to install the > full tarball to get these options? Any help greatly appreciated. They don't exist in 7.3. Why are you consulting 7.4 documentation for a 7.3 installation? regards, tom lane From pgsql-performance-owner@postgresql.org Mon Nov 3 14:42:35 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DDE9AD1B502 for ; Mon, 3 Nov 2003 18:42:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 16034-04 for ; Mon, 3 Nov 2003 14:42:03 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 0371BD1B53E for ; Mon, 3 Nov 2003 14:42:02 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3863359; Mon, 03 Nov 2003 10:42:44 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: "Lee Hughes" , Subject: Re: join_collapse_limit, from_collapse_limit options missing Date: Mon, 3 Nov 2003 10:41:04 -0800 User-Agent: KMail/1.4.3 References: <70AA9BD4C82E76418F730DF1411A624F06DF12@meta01.owens.int> In-Reply-To: <70AA9BD4C82E76418F730DF1411A624F06DF12@meta01.owens.int> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311031041.04816.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/7 X-Sequence-Number: 4567 Lee, > Hi, I'm trying to set these two options to tune performance but both > return 'not a valid option name'. Dumping the pg_settings table confirms > that they are missing. I'm using the PostgreSQL packages included with > RedHat 9 (7.3.2) and Mandrake 9.2 Beta (7.3.4). Do I need to install the > full tarball to get these options? Any help greatly appreciated. If you're working from the documentation on General Bits, please notice that both of those options are marked as "new for 7.4". -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Nov 4 06:44:36 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9CE2DD1CA63 for ; Tue, 4 Nov 2003 10:44:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 65843-08 for ; Tue, 4 Nov 2003 06:44:04 -0400 (AST) Received: from omikron.sk (cepr.nustep.sk [81.0.222.49]) by svr1.postgresql.org (Postfix) with SMTP id BB899D1B991 for ; Tue, 4 Nov 2003 06:43:59 -0400 (AST) Received: (qmail 3282 invoked from network); 4 Nov 2003 10:43:52 -0000 Received: from nustep.casablanca.sk (HELO stratos) (81.0.208.19) by 0 with SMTP; 4 Nov 2003 10:43:52 -0000 Message-ID: <00e201c3a2c0$89409f00$0200a8c0@stratos> From: "Cestmir Hybl" To: References: <017201c39d85$533eed70$0200a8c0@stratos> Subject: Re: Ignoring index on (A is null), (A is not null) conditions Date: Tue, 4 Nov 2003 11:43:51 +0100 Organization: NUSTEP s.r.o. MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-2" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests=REFERENCES X-Spam-Level: X-Archive-Number: 200311/8 X-Sequence-Number: 4568 First of all, thanks for all your suggestions. They were of two classes: 1. use different data representation (special constant from column domain instead of NULL) This is possible, of course, but it makes data model less portable and requires changes in database abstraction layer of application. 2. use partial indexes This is suitable for single null-allowed column index. With increasing number of null-allowed columns inside index, the number of partial indexes required grows exponentially. All RDBMSs I ever used (Sybase, MSSQL, or even MySQL) were using index to filter by expressions containing is NULL conditions /(A is NULL), (A is not NULL), (A is NULL or A = const), (A is NULL or A > const)/ so it seems pretty strange to me that PostgreSQL does not. Is this sheduled feature at least? CH From pgsql-performance-owner@postgresql.org Tue Nov 4 10:50:08 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8C00AD1B50D for ; Tue, 4 Nov 2003 14:50:06 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 12573-02 for ; Tue, 4 Nov 2003 10:49:39 -0400 (AST) Received: from smtp1.nodak.edu (smtp1.NoDak.edu [134.129.111.50]) by svr1.postgresql.org (Postfix) with ESMTP id 4415CD1B95A for ; Tue, 4 Nov 2003 10:49:35 -0400 (AST) Received: from ndsu.nodak.edu (webmail1.ndsu.NoDak.edu [134.129.111.141]) by smtp1.nodak.edu (8.11.6/8.11.6) with SMTP id hA4EndB17845 for ; Tue, 4 Nov 2003 08:49:39 -0600 Received: from 134.129.92.208 (SquirrelMail authenticated user radha.manohar) by webmail.ndsu.nodak.edu with HTTP; Tue, 4 Nov 2003 08:49:39 -0600 (CST) Message-ID: <1243.134.129.92.208.1067957379.squirrel@webmail.ndsu.nodak.edu> Date: Tue, 4 Nov 2003 08:49:39 -0600 (CST) Subject: Re: Response time From: To: X-Priority: 3 Importance: Normal X-Mailer: SquirrelMail (version 1.2.11) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=1.1 tagged_above=0.0 required=5.0 tests=NO_REAL_NAME X-Spam-Level: * X-Archive-Number: 200311/9 X-Sequence-Number: 4569 How do we measure the response time in postgresql? Your response would be very much appreciated. Thanks and Regards, Radha From pgsql-performance-owner@postgresql.org Tue Nov 4 11:02:56 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6BEF5D1B53B for ; Tue, 4 Nov 2003 15:02:47 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 12573-04 for ; Tue, 4 Nov 2003 11:02:20 -0400 (AST) Received: from kix.fsv.cvut.cz (Kix.FSV.CVUT.CZ [147.32.129.84]) by svr1.postgresql.org (Postfix) with ESMTP id 827C9D1CA66 for ; Tue, 4 Nov 2003 11:02:15 -0400 (AST) Received: from localhost (stehule@localhost) by kix.fsv.cvut.cz (8.11.6/8.11.6) with ESMTP id hA4F2Bj13097; Tue, 4 Nov 2003 16:02:11 +0100 Date: Tue, 4 Nov 2003 16:02:11 +0100 (CET) From: Pavel Stehule To: radha.manohar@ndsu.nodak.edu Cc: pgsql-performance@postgresql.org Subject: Re: Response time In-Reply-To: <1243.134.129.92.208.1067957379.squirrel@webmail.ndsu.nodak.edu> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/10 X-Sequence-Number: 4570 Hello explain analyse select * from lidi; QUERY PLAN ----------------------------------------------------------------------------------------------- Seq Scan on lidi (cost=0.00..1.04 rows=4 width=96) (actual time=0.046..0.092 rows=4 loops=1) Total runtime: 0.369 ms Regards Pavel On Tue, 4 Nov 2003 radha.manohar@ndsu.nodak.edu wrote: > How do we measure the response time in postgresql? > > Your response would be very much appreciated. > > Thanks and Regards, > > Radha > > > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster > From pgsql-performance-owner@postgresql.org Tue Nov 4 16:13:10 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 85250D1B518 for ; Tue, 4 Nov 2003 20:13:09 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 64568-09 for ; Tue, 4 Nov 2003 16:12:39 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id A7921D1B8DA for ; Tue, 4 Nov 2003 16:12:37 -0400 (AST) Received: from 6-allhosts (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id BAEA71D91; Tue, 4 Nov 2003 15:12:37 -0500 (EST) Subject: Re: Response time From: Neil Conway To: radha.manohar@ndsu.nodak.edu Cc: PostgreSQL Performance In-Reply-To: <1243.134.129.92.208.1067957379.squirrel@webmail.ndsu.nodak.edu> References: <1243.134.129.92.208.1067957379.squirrel@webmail.ndsu.nodak.edu> Content-Type: text/plain Message-Id: <1067976756.336.0.camel@tokyo> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Tue, 04 Nov 2003 15:12:37 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/11 X-Sequence-Number: 4571 On Tue, 2003-11-04 at 09:49, radha.manohar@ndsu.nodak.edu wrote: > How do we measure the response time in postgresql? In addition to EXPLAIN ANALYZE, the log_min_duration_statement configuration variable and the \timing psql command might also be useful. -Neil From pgsql-performance-owner@postgresql.org Wed Nov 5 09:01:07 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D6639D1BB4E for ; Wed, 5 Nov 2003 13:01:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 36099-04 for ; Wed, 5 Nov 2003 09:00:35 -0400 (AST) Received: from main.gmane.org (main.gmane.org [80.91.224.249]) by svr1.postgresql.org (Postfix) with ESMTP id 399D5D1BB7A for ; Wed, 5 Nov 2003 09:00:31 -0400 (AST) Received: from root by main.gmane.org with local (Exim 3.35 #1 (Debian)) id 1AHNGz-0006Fw-00 for ; Wed, 05 Nov 2003 14:00:33 +0100 X-Injected-Via-Gmane: http://gmane.org/ To: pgsql-performance@postgresql.org Received: from sea.gmane.org ([80.91.224.252]) by main.gmane.org with esmtp (Exim 3.35 #1 (Debian)) id 1AHKxE-0002zy-00 for ; Wed, 05 Nov 2003 11:32:00 +0100 Received: from news by sea.gmane.org with local (Exim 3.35 #1 (Debian)) id 1AHKxE-0002JI-00 for ; Wed, 05 Nov 2003 11:32:00 +0100 From: Paul Ganainm Subject: Interbase/Firebird - any users out there - what's the performance like compared to PostgreSQL? Date: Wed, 5 Nov 2003 10:28:25 -0000 Organization: N�ant Lines: 42 Message-ID: X-Complaints-To: usenet@sea.gmane.org X-Newsreader: MicroPlanet Gravity v2.50 X-Virus-Scanned: by amavisd-new at postgresql.org X-Amavis-Alert: BAD HEADER Non-encoded 8-bit data (char E9 hex) in message header 'Organization' Organization: N\351ant\n ^ X-Archive-Number: 200311/12 X-Sequence-Number: 4572 Hi all, I have just started with PostgreSQL on Linux and in the past I've done a good bit of work on Interbase (both on Windows and Linux). What I want to know here is What do people here think of Interbase/Firebird? Has anybody done performance metrics or could they point me to a comparison between the two? Does Interbase/Firebird have (as far as people here are concerned) any show-stoppers in terms of functionality which they do have on PostgreSQL? Or, indeed, the other way round? I'm not interested in starting a flame war or anything like that - just a presentation of the facts as you see them, and if you want to put in your opinion also, that's fine, just make a note! TIA. Paul... -- plinehan__AT__yahoo__DOT__com C++ Builder 5 SP1, Interbase 6.0.1.6 IBX 5.04 W2K Pro Please do not top-post. From pgsql-performance-owner@postgresql.org Wed Nov 5 09:16:29 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E4559D1B4E6 for ; Wed, 5 Nov 2003 13:16:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 42526-01 for ; Wed, 5 Nov 2003 09:15:44 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 154E6D1BB4E for ; Wed, 5 Nov 2003 09:15:41 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Wed, 05 Nov 2003 06:15:47 -0700 Message-ID: <3FA8F7F8.1070602@myrealbox.com> Date: Wed, 05 Nov 2003 18:45:36 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Paul Ganainm Cc: pgsql-performance@postgresql.org Subject: Re: Interbase/Firebird - any users out there - what's the References: In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/13 X-Sequence-Number: 4573 Paul Ganainm wrote: > Does Interbase/Firebird have (as far as people here are concerned) any > show-stoppers in terms of functionality which they do have on > PostgreSQL? Or, indeed, the other way round? Personally I think native windows port is plus that interbase/firebird has over postgresql. It also has native threaded model which *could* be benefitial at times. I combed docs once or twice. It looks like the linux/unix support is failry new so I don't know how it stacks in production and performance. OTOH, I don't like storing entire database in one file. That could get messy. Otherwise two databases are on par, at least on paper. I hope SQL compliance of interbase/firebird is better than mysql and closer to postgresql, +/- delta. Of course PG has it's own goodies. Rules/Create language are just tip of iceberg. Can you come up with some relative comparison? We can help you from postgresql side..:-) That would be great. Bye Shridhar From pgsql-performance-owner@postgresql.org Wed Nov 5 10:06:28 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 13FEED1B95A for ; Wed, 5 Nov 2003 14:06:25 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45262-03 for ; Wed, 5 Nov 2003 10:05:56 -0400 (AST) Received: from lucifer.oficina (unknown [200.69.203.237]) by svr1.postgresql.org (Postfix) with ESMTP id 20476D1BB7E for ; Wed, 5 Nov 2003 10:05:51 -0400 (AST) Received: from [192.168.1.190] (taz.oficina [192.168.1.190]) (authenticated bits=0) by lucifer.oficina (8.12.9/8.12.9) with ESMTP id hA5E5lAN099770; Wed, 5 Nov 2003 11:05:47 -0300 (ART) (envelope-from franco@akyasociados.com.ar) Subject: Re: Interbase/Firebird - any users out there - what's From: Franco Bruno Borghesi To: Paul Ganainm Cc: pgsql-performance@postgresql.org In-Reply-To: References: Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-pawMCHeqdTzgT2a8D//u" Organization: AK y Asociados S.R.L. Message-Id: <1068041145.783.10.camel@taz> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.4 Date: Wed, 05 Nov 2003 11:05:46 -0300 X-MailScanner-Information: Please contact the ISP for more information X-MailScanner: Found to be clean X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/14 X-Sequence-Number: 4574 --=-pawMCHeqdTzgT2a8D//u Content-Type: multipart/alternative; boundary="=-zV7wHhfKmlh1MGvarLo9" --=-zV7wHhfKmlh1MGvarLo9 Content-Type: text/plain Content-Transfer-Encoding: quoted-printable About a year ago I programmed a php/firebird application, and I've never had a problem with firebird. It's a small database (a few megabytes), but it just works day after day. I've seen firebird has updatable views, they seem to work very well. I have the feeling that It's not as flexible as postgresql, but I still like it. It would be my choice for win32 applications (until win32 postgresql port become available). On Wed, 2003-11-05 at 07:28, Paul Ganainm wrote: > Hi all, >=20 >=20 > I have just started with PostgreSQL on Linux and in the past I've done a= =20 > good bit of work on Interbase (both on Windows and Linux). >=20 >=20 > What I want to know here is >=20 >=20 > What do people here think of Interbase/Firebird? >=20 >=20 > Has anybody done performance metrics or could they point me to a=20 > comparison between the two? >=20 >=20 > Does Interbase/Firebird have (as far as people here are concerned) any=20 > show-stoppers in terms of functionality which they do have on=20 > PostgreSQL? Or, indeed, the other way round? >=20 >=20 > I'm not interested in starting a flame war or anything like that - just= =20 > a presentation of the facts as you see them, and if you want to put in=20 > your opinion also, that's fine, just make a note! >=20 >=20 > TIA. >=20 >=20 > Paul... >=20 --=-zV7wHhfKmlh1MGvarLo9 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable About a year ago I programmed a php/firebird application, and I've never ha= d a problem with firebird. It's a small database (a few megabytes), but it = just works day after day.

I've seen firebird has updatable views, they seem to work very well.

I have the feeling that It's not as flexible as postgresql, but I still lik= e it. It would be my choice for win32 applications (until win32 postgresql = port become available).


On Wed, 2003-11-05 at 07:28, Paul Ganainm wrote: 
Hi all,


I have just started with PostgreSQL on Linux and in the past I've done a=20
good bit of work on Interbase (both on Windows and Linux).


What I want to know here is


What do people here think of Interbase/Firebird?


Has anybody done performance metrics or could they point me to a=20
comparison between the two?


Does Interbase/Firebird have (as far as people here are concerned) any=20
show-stoppers in terms of functionality which they do have on=20
PostgreSQL? Or, indeed, the other way round?


I'm not interested in starting a flame war or anything like that - just=20
a presentation of the facts as you see them, and if you want to put in=20
your opinion also, that's fine, just make a note!


TIA.


Paul...
--=-zV7wHhfKmlh1MGvarLo9-- --=-pawMCHeqdTzgT2a8D//u Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2 (FreeBSD) iD8DBQA/qQO521dVnhLsBV0RAsAtAJ9fSlVl2HhXpZy997SM1CEAKx8TOgCfYlHx 6Qjmzj3zKiiRN4Oe1LkkHjE= =U6qz -----END PGP SIGNATURE----- --=-pawMCHeqdTzgT2a8D//u-- From pgsql-performance-owner@postgresql.org Wed Nov 5 13:36:13 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E1574D1B4F4 for ; Wed, 5 Nov 2003 17:36:11 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 83418-05 for ; Wed, 5 Nov 2003 13:35:41 -0400 (AST) Received: from smtp1.nodak.edu (smtp1.NoDak.edu [134.129.111.50]) by svr1.postgresql.org (Postfix) with ESMTP id 0580FD1B4ED for ; Wed, 5 Nov 2003 13:35:40 -0400 (AST) Received: from ndsu.nodak.edu (webmail1.ndsu.NoDak.edu [134.129.111.141]) by smtp1.nodak.edu (8.11.6/8.11.6) with SMTP id hA5HZHB21072; Wed, 5 Nov 2003 11:35:20 -0600 Received: from 134.129.92.208 (SquirrelMail authenticated user radha.manohar) by webmail.ndsu.nodak.edu with HTTP; Wed, 5 Nov 2003 11:35:22 -0600 (CST) Message-ID: <1706.134.129.92.208.1068053722.squirrel@webmail.ndsu.nodak.edu> Date: Wed, 5 Nov 2003 11:35:22 -0600 (CST) Subject: Re: Response time From: To: In-Reply-To: <1067976756.336.0.camel@tokyo> References: <1243.134.129.92.208.1067957379.squirrel@webmail.ndsu.nodak.edu> <1067976756.336.0.camel@tokyo> X-Priority: 3 Importance: Normal Cc: X-Mailer: SquirrelMail (version 1.2.11) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/15 X-Sequence-Number: 4575 The \timing psql command gives different time for the same query executed repeatedly. So, how can we know the exact response time for any query? Thanks and Regards, Radha > On Tue, 2003-11-04 at 09:49, radha.manohar@ndsu.nodak.edu wrote: >> How do we measure the response time in postgresql? > > In addition to EXPLAIN ANALYZE, the log_min_duration_statement > configuration variable and the \timing psql command might also be > useful. > > -Neil > > > > ---------------------------(end of broadcast)--------------------------- > TIP 8: explain analyze is your friend From pgsql-performance-owner@postgresql.org Wed Nov 5 13:41:00 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 7C0CDD1BB9A for ; Wed, 5 Nov 2003 17:40:58 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 85177-01 for ; Wed, 5 Nov 2003 13:40:28 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id 51429D1BB7B for ; Wed, 5 Nov 2003 13:40:27 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 6AB381DF2; Wed, 5 Nov 2003 12:40:20 -0500 (EST) To: Cc: Subject: Re: Response time From: Neil Conway In-Reply-To: <1706.134.129.92.208.1068053722.squirrel@webmail.ndsu.nodak.edu> (radha manohar's message of "Wed, 5 Nov 2003 11:35:22 -0600 (CST)") References: <1243.134.129.92.208.1067957379.squirrel@webmail.ndsu.nodak.edu> <1067976756.336.0.camel@tokyo> <1706.134.129.92.208.1068053722.squirrel@webmail.ndsu.nodak.edu> Date: Wed, 05 Nov 2003 12:40:18 -0500 Message-ID: <87ad7ar9x9.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/16 X-Sequence-Number: 4576 writes: > The \timing psql command gives different time for the same query executed > repeatedly. That's probably because executing the query repeatedly results in different execution times, as one would expect. \timing returns the "exact" query response time, nevertheless. -Neil From pgsql-performance-owner@postgresql.org Wed Nov 5 14:04:54 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 86020D1B947 for ; Wed, 5 Nov 2003 18:04:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 84183-10 for ; Wed, 5 Nov 2003 14:04:23 -0400 (AST) Received: from mail.libertyrms.com (unknown [209.167.124.227]) by svr1.postgresql.org (Postfix) with ESMTP id 609B0D1B8DF for ; Wed, 5 Nov 2003 14:04:22 -0400 (AST) Received: from [10.1.2.130] (helo=dba2) by mail.libertyrms.com with esmtp (Exim 3.22 #3 (Debian)) id 1AHS10-0004Wp-00 for ; Wed, 05 Nov 2003 13:04:22 -0500 Received: by dba2 (Postfix, from userid 1019) id C8468C9B3; Wed, 5 Nov 2003 13:04:22 -0500 (EST) Date: Wed, 5 Nov 2003 13:04:22 -0500 From: Andrew Sullivan To: pgsql-performance@postgresql.org Subject: Re: Response time Message-ID: <20031105180422.GF4506@libertyrms.info> Mail-Followup-To: Andrew Sullivan , pgsql-performance@postgresql.org References: <1243.134.129.92.208.1067957379.squirrel@webmail.ndsu.nodak.edu> <1067976756.336.0.camel@tokyo> <1706.134.129.92.208.1068053722.squirrel@webmail.ndsu.nodak.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1706.134.129.92.208.1068053722.squirrel@webmail.ndsu.nodak.edu> User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/17 X-Sequence-Number: 4577 On Wed, Nov 05, 2003 at 11:35:22AM -0600, radha.manohar@ndsu.nodak.edu wrote: > The \timing psql command gives different time for the same query executed > repeatedly. Why do you believe that the same query will always take the same time to execute? A -- ---- Andrew Sullivan 204-4141 Yonge Street Afilias Canada Toronto, Ontario Canada M2P 2A8 +1 416 646 3304 x110 From pgsql-performance-owner@postgresql.org Thu Nov 6 04:44:02 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CB326D1B8DF for ; Thu, 6 Nov 2003 08:44:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 33882-03 for ; Thu, 6 Nov 2003 04:43:31 -0400 (AST) Received: from sys-srv-1.xtec.ro (xtec.b.astral.ro [213.164.250.17]) by svr1.postgresql.org (Postfix) with ESMTP id D3D59D1BB59 for ; Thu, 6 Nov 2003 04:43:28 -0400 (AST) Received: from fx.ro (xtro-ws-1.xtec.ro [10.60.4.1]) by sys-srv-1.xtec.ro (Postfix) with ESMTP id D03EB3FE7 for ; Thu, 6 Nov 2003 10:42:40 +0200 (EET) Message-ID: <3FAA09BB.6050705@fx.ro> Date: Thu, 06 Nov 2003 10:43:39 +0200 From: Adrian Maier User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5b) Gecko/20030901 Thunderbird/0.2 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: horrible performance when trying to connect to PostgreSQL using ODBC Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/18 X-Sequence-Number: 4578 Hello guys ! I am trying to connect to pgsql using ODBC from Visual Objects (for the ones who don't know what is Visual Objects, it's an IDE with its own object-oriented language made by Computer Associates. VO's roots are related to Clipper. ). The time it takes to get the list of the tables, and then to get the data and display it way too long. It took more than 16 seconds to connect, get 200 records and print them on the console. I know for sure that it's not pgsql's fault: with a small c program that uses libpq i got the same data instantaneously. Also, it looks like it's not VO's fault (i know that there are people who used VO against big databases). So, the problem must be related to the ODBC (the pgsql driver or something else?). Anyway, i've tried to open the same table from MSAccess, and the performance was even worse. Just an example: for going from one record to the next one it lasts about 4-5 seconds. The machine is a Pentium2 266mhz (128mb) laptop running windows NT. Today, I have tried to access pgsql from MSAccess at work. It worked very fast. I've used the same ODBC drivers, but the machine is faster (p4 2.4ghz), has more memory, and the OS is windows 2000. But it is hard for me to believe that a Pentium2 machine with 128mb RAM can't handle 200 records ! I begin to think that there is something fishy about the ODBC support installed on the laptop (i have recently "inherited" the laptop from someone else, and i haven't reinstalled the os ). What do you guys think? what could be done to make things go faster? I have also two more questions. After installing the pgsql odbc drivers, there are 3 drivers that appear: "PostgreSQL", "POstgreSQL legacy" and "PostgreSQL unicode". What's the difference between "PostgreSQL" and "PostgreSQL lecacy"? Are there any options of the odbc driver that could cause a performance boost? Thanks you, Adrian Maier From pgsql-performance-owner@postgresql.org Thu Nov 6 10:58:27 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 98D98D1BACF for ; Thu, 6 Nov 2003 14:58:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 90254-02 for ; Thu, 6 Nov 2003 10:57:57 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 378F7D1BAE3 for ; Thu, 6 Nov 2003 10:57:53 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hA6Evp19019450; Thu, 6 Nov 2003 09:57:51 -0500 (EST) To: Adrian Maier Cc: pgsql-performance@postgresql.org Subject: Re: horrible performance when trying to connect to PostgreSQL using ODBC In-reply-to: <3FAA09BB.6050705@fx.ro> References: <3FAA09BB.6050705@fx.ro> Comments: In-reply-to Adrian Maier message dated "Thu, 06 Nov 2003 10:43:39 +0200" Date: Thu, 06 Nov 2003 09:57:51 -0500 Message-ID: <19449.1068130671@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/19 X-Sequence-Number: 4579 Adrian Maier writes: > I begin to think that there is something fishy about the ODBC > support installed on the laptop (i have recently "inherited" the > laptop from someone else, and i haven't reinstalled the os ). You'd be better off asking these questions on pgsql-odbc. I'm not sure that any of the ODBC gurus read this list. I believe that the ODBC driver has some incredibly extensive, and expensive, debug-logging options. I dunno if max logging would entirely explain the slowdown you see, but definitely check what you have turned on. regards, tom lane From pgsql-performance-owner@postgresql.org Sat Nov 8 16:24:17 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B6255D1B99F for ; Thu, 6 Nov 2003 17:17:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 10032-09 for ; Thu, 6 Nov 2003 13:17:16 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 3FE56D1B4E6 for ; Thu, 6 Nov 2003 13:17:16 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hA6HHFNu004498 for ; Thu, 6 Nov 2003 17:17:15 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hA6GvHp9000404 for pgsql-performance@postgresql.org; Thu, 6 Nov 2003 16:57:17 GMT From: "Private" X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Interbase/Firebird - any users out there - what's the performance like compared to PostgreSQL? Date: Thu, 6 Nov 2003 11:57:15 -0500 Organization: Hub.Org Networking Services Lines: 65 Message-ID: References: X-Complaints-To: usenet@news.hub.org X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests=PRIORITY_NO_NAME, QUOTED_EMAIL_TEXT, REFERENCES X-Spam-Level: X-Archive-Number: 200311/23 X-Sequence-Number: 4583 Try this benchmark on PostgreSQL, MySQL, FireBird, Oracle: http://go.jitbot.com/dbbench-pg-fb-mys-orcl This page expired but if you click on "Cache" on the side of JitBot search pane, you'll see the cached page: http://go.jitbot.com/opensource-dbs-table Cheers, "Paul Ganainm" wrote in message news:MPG.1a12e0e45193bd1b98969a@news.gmane.org... > > > Hi all, > > > I have just started with PostgreSQL on Linux and in the past I've done a > good bit of work on Interbase (both on Windows and Linux). > > > What I want to know here is > > > What do people here think of Interbase/Firebird? > > > Has anybody done performance metrics or could they point me to a > comparison between the two? > > > Does Interbase/Firebird have (as far as people here are concerned) any > show-stoppers in terms of functionality which they do have on > PostgreSQL? Or, indeed, the other way round? > > > I'm not interested in starting a flame war or anything like that - just > a presentation of the facts as you see them, and if you want to put in > your opinion also, that's fine, just make a note! > > > TIA. > > > Paul... > > > -- > > plinehan__AT__yahoo__DOT__com > > C++ Builder 5 SP1, Interbase 6.0.1.6 IBX 5.04 W2K Pro > > Please do not top-post. > > > ---------------------------(end of broadcast)--------------------------- > TIP 2: you can get off all lists at once with the unregister command > (send "unregister YourEmailAddressHere" to majordomo@postgresql.org) > From pgsql-performance-owner@postgresql.org Thu Nov 6 15:17:50 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E0408D1B564 for ; Thu, 6 Nov 2003 19:17:47 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 36146-04 for ; Thu, 6 Nov 2003 15:17:17 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id A6221D1B522 for ; Thu, 6 Nov 2003 15:17:16 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hA6JHGNu030507 for ; Thu, 6 Nov 2003 19:17:16 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hA6JA2wW028824 for pgsql-performance@postgresql.org; Thu, 6 Nov 2003 19:10:02 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Interbase/Firebird - any users out there - what's the performance like compared to PostgreSQL? Date: Thu, 06 Nov 2003 13:23:06 -0500 Organization: Hub.Org Networking Services Lines: 18 Message-ID: <60n0b9uzjp.fsf@dev6.int.libertyrms.info> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org User-Agent: Gnus/5.1003 (Gnus v5.10.3) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:YytxfyG8bBSb27zjYXazEqpCb90= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/20 X-Sequence-Number: 4580 "Private" writes: > Try this benchmark on PostgreSQL, MySQL, FireBird, Oracle: > > http://go.jitbot.com/dbbench-pg-fb-mys-orcl It looks like a good candidate for adding in a plpgsql stored procedure to get similar speedups to what was gotten with the Oracle benchmark. It looks, from the numbers, as though Firebird is _slightly_ slower than PostgreSQL, which seems not totally remarkable. I'd expect that reimplementing the benchmark inside a stored procedure would provide similar improvements with Firebird... -- (reverse (concatenate 'string "ofni.smrytrebil" "@" "enworbbc")) Christopher Browne (416) 646 3304 x124 (land) From pgsql-performance-owner@postgresql.org Thu Nov 6 15:54:58 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 38BB9D1B557 for ; Thu, 6 Nov 2003 19:54:56 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 41709-07 for ; Thu, 6 Nov 2003 15:54:26 -0400 (AST) Received: from fuji.krosing.net (217-159-136-226-dsl.kt.estpak.ee [217.159.136.226]) by svr1.postgresql.org (Postfix) with ESMTP id 35432D1B4E1 for ; Thu, 6 Nov 2003 15:54:24 -0400 (AST) Received: from fuji.krosing.net (localhost.localdomain [127.0.0.1]) by fuji.krosing.net (8.12.8/8.12.8) with ESMTP id hA6JsFgT003887; Thu, 6 Nov 2003 21:54:17 +0200 Received: (from hannu@localhost) by fuji.krosing.net (8.12.8/8.12.8/Submit) id hA6JsDe1003885; Thu, 6 Nov 2003 21:54:13 +0200 X-Authentication-Warning: fuji.krosing.net: hannu set sender to hannu@tm.ee using -f Subject: Re: Interbase/Firebird - any users out there - what's From: Hannu Krosing To: Christopher Browne Cc: pgsql-performance@postgresql.org In-Reply-To: <60n0b9uzjp.fsf@dev6.int.libertyrms.info> References: <60n0b9uzjp.fsf@dev6.int.libertyrms.info> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1068148452.3735.3.camel@fuji.krosing.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Thu, 06 Nov 2003 21:54:13 +0200 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/21 X-Sequence-Number: 4581 Christopher Browne kirjutas N, 06.11.2003 kell 20:23: > "Private" writes: > > Try this benchmark on PostgreSQL, MySQL, FireBird, Oracle: > > > > http://go.jitbot.com/dbbench-pg-fb-mys-orcl > > It looks like a good candidate for adding in a plpgsql stored > procedure to get similar speedups to what was gotten with the Oracle > benchmark. It would also be interesting to see the same test run on Postgresql on Linux/UNIX. PgSQL on win2000 (most likely using cygwin) is probably not the best you can get out of that hardware. ----------- Hannu From pgsql-performance-owner@postgresql.org Thu Nov 6 18:59:28 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6AA6ED1B523 for ; Thu, 6 Nov 2003 22:59:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 73572-02 for ; Thu, 6 Nov 2003 18:58:58 -0400 (AST) Received: from candle.pha.pa.us (unknown [207.106.42.251]) by svr1.postgresql.org (Postfix) with ESMTP id 79CBDD1BB5D for ; Thu, 6 Nov 2003 18:58:54 -0400 (AST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.11.6) id hA6Mwq504482; Thu, 6 Nov 2003 17:58:52 -0500 (EST) From: Bruce Momjian Message-Id: <200311062258.hA6Mwq504482@candle.pha.pa.us> Subject: Re: Pg+Linux swap use In-Reply-To: <87y8uy91ds.fsf@stark.dyndns.tv> To: Greg Stark Date: Thu, 6 Nov 2003 17:58:52 -0500 (EST) Cc: William Yu , pgsql-performance@postgresql.org X-Mailer: ELM [version 2.4ME+ PL108 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/22 X-Sequence-Number: 4582 Greg Stark wrote: > > William Yu writes: > > > Rob Sell wrote: > > > > > Not being one to hijack threads, but I haven't heard of this performance hit > > > when using HT, I have what should all rights be a pretty fast server, dual > > > 2.4 Xeons with HT 205gb raid 5 array, 1 gig of memory. And it is only 50% as > > > fast as my old server which was a dual AMD MP 1400's with a 45gb raid 5 > > > array and 1gb of ram. > > > > Not to get into a big Intel vs AMD argument but 50% sounds about right. Let's > > first assume that the QS rating for the MP1400 is relatively accurate and > > convert that to a 1.4GHz Xeon. 2.4/1.4 = +71%. Since processor performance > > does not increase linearly with clockspeed, 50% is in line with expectations. > > Hm. You've read "50% as fast" as "50% faster". > I wonder which the original poster intended. Hyper-threading makes 2 cpus be 4 cpu's, but the 4 cpu's are each only 70% as fast, so HT is taking 2x cpus and making it 4x0.70 cpu's, which gives 2.80 cpu's, and you get that only if you are hammering all four cpu's with a full load. Imagine ifd get two cpu-bound processes on the first die (first 2 cpu's of 4) and the other CPU die is idle, and you can see that HT isn't all that useful unless you are sure to keep all 4 cpu's busy. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073 From pgsql-novice-owner@postgresql.org Sun Nov 9 11:06:50 2003 X-Original-To: pgsql-novice-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id ABFA0D1B50F; Sun, 9 Nov 2003 15:06:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 84862-07; Sun, 9 Nov 2003 11:06:16 -0400 (AST) Received: from smtp1.nodak.edu (smtp1.NoDak.edu [134.129.111.50]) by svr1.postgresql.org (Postfix) with ESMTP id C7B1AD1B4EF; Sun, 9 Nov 2003 11:06:10 -0400 (AST) Received: from ndsu.nodak.edu (webmail1.ndsu.NoDak.edu [134.129.111.141]) by smtp1.nodak.edu (8.11.6/8.11.6) with SMTP id hA9F6EB24003; Sun, 9 Nov 2003 09:06:14 -0600 Received: from 134.129.110.96 (SquirrelMail authenticated user radha.manohar) by webmail.ndsu.nodak.edu with HTTP; Sun, 9 Nov 2003 09:06:14 -0600 (CST) Message-ID: <3324.134.129.110.96.1068390374.squirrel@webmail.ndsu.nodak.edu> Date: Sun, 9 Nov 2003 09:06:14 -0600 (CST) Subject: Re: error while executing a c program with embedded sql From: To: X-Priority: 3 Importance: Normal Cc: X-Mailer: SquirrelMail (version 1.2.11) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/60 X-Sequence-Number: 8763 I have a c program called test1.pgc with some sql statements embedded in it. The program was preprocessed, compiled and linked. Now, I have the executable test1. When I run the executable it says, ./test1: error while loading shared libraries: libecpg.so.3: cannot open shared object file: No such file or directory What does it mean by this error message? What should I do to correct this error and run the executable successfully? Your response would be very much appreciated. Thanks and Regards, Radha From pgsql-novice-owner@postgresql.org Sun Nov 9 11:34:42 2003 X-Original-To: pgsql-novice-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A43C3D1B4EF; Sun, 9 Nov 2003 15:34:35 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 00602-02; Sun, 9 Nov 2003 11:34:07 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 90A37D1B4E6; Sun, 9 Nov 2003 11:34:03 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hA9FY719003364; Sun, 9 Nov 2003 10:34:07 -0500 (EST) To: radha.manohar@ndsu.nodak.edu Cc: pgsql-novice@postgresql.org, pgsql-performance@postgresql.org Subject: Re: error while executing a c program with embedded sql In-reply-to: <3324.134.129.110.96.1068390374.squirrel@webmail.ndsu.nodak.edu> References: <3324.134.129.110.96.1068390374.squirrel@webmail.ndsu.nodak.edu> Comments: In-reply-to message dated "Sun, 09 Nov 2003 09:06:14 -0600" Date: Sun, 09 Nov 2003 10:34:06 -0500 Message-ID: <3363.1068392046@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/61 X-Sequence-Number: 8764 writes: > ./test1: error while loading shared libraries: libecpg.so.3: cannot open > shared object file: No such file or directory The dynamic linker is failing to find either libecpg.so itself, or one of the shared libraries it depends on (perhaps libpq.so). If you are on a Linux system you probably want to fix your ldconfig configuration so that all these libraries are found automatically. You can run ldd on a particular executable or shared library to see what libraries it references and whether those libraries are getting found in the proper places. Note that it is entirely possible for the program linking stage to succeed but dynamic linking to fail at runtime. For various reasons the search rules are not quite the same in the two contexts ... regards, tom lane From pgsql-novice-owner@postgresql.org Sun Nov 9 13:36:45 2003 X-Original-To: pgsql-novice-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 754E1D1B506 for ; Sun, 9 Nov 2003 17:36:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 02715-08 for ; Sun, 9 Nov 2003 13:36:12 -0400 (AST) Received: from anchor-post-35.mail.demon.net (anchor-post-35.mail.demon.net [194.217.242.85]) by svr1.postgresql.org (Postfix) with ESMTP id C0EFFD1B509 for ; Sun, 9 Nov 2003 13:36:10 -0400 (AST) Received: from lfix.demon.co.uk ([80.177.205.209] helo=cerberus.lfix.co.uk) by anchor-post-35.mail.demon.net with esmtp (Exim 3.35 #1) id 1AItTu-000KRu-0Z; Sun, 09 Nov 2003 17:36:10 +0000 Received: from linda.lfix.co.uk ([192.168.1.1]) by cerberus.lfix.co.uk with esmtp (Exim 4.24) id 1AItQl-00082b-MO; Sun, 09 Nov 2003 17:32:55 +0000 Subject: Re: error while executing a c program with embedded sql From: Oliver Elphick To: radha.manohar@ndsu.nodak.edu Cc: pgsql-novice@postgresql.org In-Reply-To: <3324.134.129.110.96.1068390374.squirrel@webmail.ndsu.nodak.edu> References: <3324.134.129.110.96.1068390374.squirrel@webmail.ndsu.nodak.edu> Content-Type: text/plain Organization: LFIX Limited Message-Id: <1068399145.2175.7.camel@linda.lfix.co.uk> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Sun, 09 Nov 2003 17:32:25 +0000 Content-Transfer-Encoding: 7bit X-LFIX-MailScanner: Found to be clean X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/62 X-Sequence-Number: 8765 On Sun, 2003-11-09 at 15:06, radha.manohar@ndsu.nodak.edu wrote: > I have a c program called test1.pgc with some sql statements embedded in > it. The program was preprocessed, compiled and linked. Now, I have the > executable test1. > > When I run the executable it says, > > ./test1: error while loading shared libraries: libecpg.so.3: cannot open > shared object file: No such file or directory > > What does it mean by this error message? What should I do to correct this > error and run the executable successfully? Shared libraries are loaded from directories specified to the system by ldconfig. Your shared library, libecpg.so.3, is in a PostgreSQL directory, such as /usr/local/pgsql/lib, which has not been added to the directories known to the loader. If you are able to add that directory with ldconfig, that is the best way to do it, but it requires root privilege. Otherwise you can set the environment variable LD_LIBRARY_PATH, thus: export LD_LIBRARY_PATH=/usr/local/pgsql/lib before you run the program, or you can use LD_PRELOAD: LD_PRELOAD=/usr/local/pgsql/lib/libecpg.so.3 ./test1 -- Oliver Elphick Oliver.Elphick@lfix.co.uk Isle of Wight, UK http://www.lfix.co.uk/oliver GPG: 1024D/3E1D0C1C: CA12 09E0 E8D5 8870 5839 932A 614D 4C34 3E1D 0C1C ======================================== "O death, where is thy sting? O grave, where is thy victory?" 1 Corinthians 15:55 From pgsql-novice-owner@postgresql.org Mon Nov 10 11:01:55 2003 X-Original-To: pgsql-novice-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EDE19D1B4FE; Mon, 10 Nov 2003 15:01:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 22043-03; Mon, 10 Nov 2003 11:01:23 -0400 (AST) Received: from smtp1.nodak.edu (smtp1.NoDak.edu [134.129.111.50]) by svr1.postgresql.org (Postfix) with ESMTP id 60591D1B4E4; Mon, 10 Nov 2003 11:01:18 -0400 (AST) Received: from ndsu.nodak.edu (webmail1.ndsu.NoDak.edu [134.129.111.141]) by smtp1.nodak.edu (8.11.6/8.11.6) with SMTP id hAAF0RB30797; Mon, 10 Nov 2003 09:00:27 -0600 Received: from 134.129.92.208 (SquirrelMail authenticated user radha.manohar) by webmail.ndsu.nodak.edu with HTTP; Mon, 10 Nov 2003 09:00:27 -0600 (CST) Message-ID: <1371.134.129.92.208.1068476427.squirrel@webmail.ndsu.nodak.edu> Date: Mon, 10 Nov 2003 09:00:27 -0600 (CST) Subject: Re: error while executing a c program with embedded sql From: To: In-Reply-To: <1068399145.2175.7.camel@linda.lfix.co.uk> References: <3324.134.129.110.96.1068390374.squirrel@webmail.ndsu.nodak.edu> <1068399145.2175.7.camel@linda.lfix.co.uk> X-Priority: 3 Importance: Normal Cc: , X-Mailer: SquirrelMail (version 1.2.11) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/85 X-Sequence-Number: 8788 Thanks a lot. IT WORKED! with your suggestions. Regards, Radha > On Sun, 2003-11-09 at 15:06, radha.manohar@ndsu.nodak.edu wrote: >> I have a c program called test1.pgc with some sql statements embedded >> in it. The program was preprocessed, compiled and linked. Now, I have >> the executable test1. >> >> When I run the executable it says, >> >> ./test1: error while loading shared libraries: libecpg.so.3: cannot >> open shared object file: No such file or directory >> >> What does it mean by this error message? What should I do to correct >> this error and run the executable successfully? > > Shared libraries are loaded from directories specified to the system by > ldconfig. Your shared library, libecpg.so.3, is in a PostgreSQL > directory, such as /usr/local/pgsql/lib, which has not been added to the > directories known to the loader. > > If you are able to add that directory with ldconfig, that is the best > way to do it, but it requires root privilege. > > Otherwise you can set the environment variable LD_LIBRARY_PATH, thus: > > export LD_LIBRARY_PATH=/usr/local/pgsql/lib > > before you run the program, or you can use LD_PRELOAD: > > LD_PRELOAD=/usr/local/pgsql/lib/libecpg.so.3 ./test1 > > -- > Oliver Elphick Oliver.Elphick@lfix.co.uk > Isle of Wight, UK > http://www.lfix.co.uk/oliver GPG: 1024D/3E1D0C1C: CA12 09E0 E8D5 8870 > 5839 932A 614D 4C34 3E1D 0C1C > ======================================== > "O death, where is thy sting? O grave, where is > thy victory?" 1 Corinthians 15:55 From pgsql-novice-owner@postgresql.org Sun Nov 9 16:13:18 2003 X-Original-To: pgsql-novice-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8C3D7D1B525 for ; Sun, 9 Nov 2003 16:27:44 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 02715-04 for ; Sun, 9 Nov 2003 12:27:13 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id 25CC4D1B532 for ; Sun, 9 Nov 2003 12:27:12 -0400 (AST) Received: (qmail 18980 invoked from network); 9 Nov 2003 16:28:18 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 9 Nov 2003 16:28:18 -0000 Message-ID: <3FAFBC68.4020601@trade-india.com> Date: Mon, 10 Nov 2003 21:57:20 +0530 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: radha.manohar@ndsu.nodak.edu Cc: pgsql-novice@postgresql.org Subject: Re: [PERFORM] error while executing a c program with embedded sql References: <3324.134.129.110.96.1068390374.squirrel@webmail.ndsu.nodak.edu> In-Reply-To: <3324.134.129.110.96.1068390374.squirrel@webmail.ndsu.nodak.edu> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/68 X-Sequence-Number: 8771 radha.manohar@ndsu.nodak.edu wrote: >I have a c program called test1.pgc with some sql statements embedded in >it. The program was preprocessed, compiled and linked. Now, I have the >executable test1. > >When I run the executable it says, > >./test1: error while loading shared libraries: libecpg.so.3: cannot open >shared object file: No such file or directory > check where the so file is located. $ locate libecpg.so.3 say its /usr/local/pgsql/lib then either add the folder above to /etc/ld.so.conf and run ldconfig as root. or $ export LD_LIBRARY_PATH=/path/to/folder/containing/the/so/file $ ./test1 the above assumes u are on linux. on unix also its similar. > >What does it mean by this error message? What should I do to correct this >error and run the executable successfully? > >Your response would be very much appreciated. > >Thanks and Regards, > >Radha > > > >---------------------------(end of broadcast)--------------------------- >TIP 2: you can get off all lists at once with the unregister command > (send "unregister YourEmailAddressHere" to majordomo@postgresql.org) > > From pgsql-performance-owner@postgresql.org Mon Nov 10 16:21:35 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id F041AD1B507 for ; Mon, 10 Nov 2003 20:21:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 70902-09 for ; Mon, 10 Nov 2003 16:21:04 -0400 (AST) Received: from ganymede.hub.org (u46n208.hfx.eastlink.ca [24.222.46.208]) by svr1.postgresql.org (Postfix) with ESMTP id 48088D1B4F1 for ; Mon, 10 Nov 2003 16:21:02 -0400 (AST) Received: by ganymede.hub.org (Postfix, from userid 1000) id B42F2365D6; Mon, 10 Nov 2003 16:18:52 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by ganymede.hub.org (Postfix) with ESMTP id AABA1349C6 for ; Mon, 10 Nov 2003 16:18:52 -0400 (AST) Date: Mon, 10 Nov 2003 16:18:52 -0400 (AST) From: "Marc G. Fournier" X-X-Sender: scrappy@ganymede.hub.org To: pgsql-performance@postgresql.org Subject: *very* slow query to summarize data for a month ... Message-ID: <20031110160144.C727@ganymede.hub.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/27 X-Sequence-Number: 4587 Table structure is simple: CREATE TABLE traffic_logs ( company_id bigint, ip_id bigint, port integer, bytes bigint, runtime timestamp without time zone ); runtime is 'day of month' ... I need to summarize the month, per company, with a query as: explain analyze SELECT ts.company_id, company_name, SUM(ts.bytes) AS total_traffic FROM company c, traffic_logs ts WHERE c.company_id = ts.company_id AND month_trunc(ts.runtime) = '2003-10-01' GROUP BY company_name,ts.company_id; and the explain looks like: QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=32000.94..32083.07 rows=821 width=41) (actual time=32983.36..47586.17 rows=144 loops=1) -> Group (cost=32000.94..32062.54 rows=8213 width=41) (actual time=32957.40..42817.88 rows=462198 loops=1) -> Sort (cost=32000.94..32021.47 rows=8213 width=41) (actual time=32957.38..36261.31 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=31321.45..31466.92 rows=8213 width=41) (actual time=13983.07..22642.14 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=5.52..7.40 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=0.02..2.78 rows=352 loops=1) -> Sort (cost=31297.04..31317.57 rows=8213 width=16) (actual time=13977.49..16794.41 rows=462198 loops=1) Sort Key: ts.company_id -> Index Scan using tl_month on traffic_logs ts (cost=0.00..30763.02 rows=8213 width=16) (actual time=0.29..5562.25 rows=462198 loops=1) Index Cond: (month_trunc(runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 47587.82 msec (14 rows) the problem is that we're only taking a few months worth of data, so I don't think there is much of a way of 'improve performance' on this, but figured I'd ask quickly before I do something rash ... Note that without the month_trunc() index, the Total runtime more then doubles: QUERY PLAN -------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=39578.63..39660.76 rows=821 width=41) (actual time=87805.47..101251.35 rows=144 loops=1) -> Group (cost=39578.63..39640.23 rows=8213 width=41) (actual time=87779.56..96824.56 rows=462198 loops=1) -> Sort (cost=39578.63..39599.17 rows=8213 width=41) (actual time=87779.52..90781.48 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=38899.14..39044.62 rows=8213 width=41) (actual time=64073.98..72783.68 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=64.66..66.55 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=1.76..61.70 rows=352 loops=1) -> Sort (cost=38874.73..38895.27 rows=8213 width=16) (actual time=64009.26..66860.71 rows=462198 loops=1) Sort Key: ts.company_id -> Seq Scan on traffic_logs ts (cost=0.00..38340.72 rows=8213 width=16) (actual time=5.02..-645982.04 rows=462198 loops=1) Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 101277.17 msec (14 rows) From pgsql-performance-owner@postgresql.org Mon Nov 10 16:40:11 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B69B0D1B507; Mon, 10 Nov 2003 20:40:09 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 72096-09; Mon, 10 Nov 2003 16:39:40 -0400 (AST) Received: from fddlnint05.fds.com (fddlnint05.fds.com [208.15.91.52]) by svr1.postgresql.org (Postfix) with ESMTP id 21B69D1B50C; Mon, 10 Nov 2003 16:39:38 -0400 (AST) In-Reply-To: <20031110160144.C727@ganymede.hub.org> Subject: Re: *very* slow query to summarize data for a month ... To: scrappy@postgresql.org Cc: pgsql-performance@postgresql.org, pgsql-performance-owner@postgresql.org X-Mailer: Lotus Notes Release 6.0 September 26, 2002 Message-ID: From: "Patrick Hatcher" Date: Mon, 10 Nov 2003 12:31:08 -0800 X-MIMETrack: Serialize by Router on FDDLNINT05/FSG/SVR/FDD(Release 5.0.4 |June 8, 2000) at 11/10/2003 03:34:04 PM MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/28 X-Sequence-Number: 4588 Do you have an index on ts.bytes? Josh had suggested this and after I put it on my summed fields, I saw a speed increase. I can't remember the article was that Josh had written about index usage, but maybe he'll chime in and supply the URL for his article. hth Patrick Hatcher "Marc G. Fournier" To Sent by: pgsql-performance@postgresql.org pgsql-performance-o cc wner@postgresql.org Subject [PERFORM] *very* slow query to 11/10/2003 12:18 PM summarize data for a month ... Table structure is simple: CREATE TABLE traffic_logs ( company_id bigint, ip_id bigint, port integer, bytes bigint, runtime timestamp without time zone ); runtime is 'day of month' ... I need to summarize the month, per company, with a query as: explain analyze SELECT ts.company_id, company_name, SUM(ts.bytes) AS total_traffic FROM company c, traffic_logs ts WHERE c.company_id = ts.company_id AND month_trunc(ts.runtime) = '2003-10-01' GROUP BY company_name,ts.company_id; and the explain looks like: QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=32000.94..32083.07 rows=821 width=41) (actual time=32983.36..47586.17 rows=144 loops=1) -> Group (cost=32000.94..32062.54 rows=8213 width=41) (actual time=32957.40..42817.88 rows=462198 loops=1) -> Sort (cost=32000.94..32021.47 rows=8213 width=41) (actual time=32957.38..36261.31 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=31321.45..31466.92 rows=8213 width=41) (actual time=13983.07..22642.14 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=5.52..7.40 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=0.02..2.78 rows=352 loops=1) -> Sort (cost=31297.04..31317.57 rows=8213 width=16) (actual time=13977.49..16794.41 rows=462198 loops=1) Sort Key: ts.company_id -> Index Scan using tl_month on traffic_logs ts (cost=0.00..30763.02 rows=8213 width=16) (actual time=0.29..5562.25 rows=462198 loops=1) Index Cond: (month_trunc(runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 47587.82 msec (14 rows) the problem is that we're only taking a few months worth of data, so I don't think there is much of a way of 'improve performance' on this, but figured I'd ask quickly before I do something rash ... Note that without the month_trunc() index, the Total runtime more then doubles: QUERY PLAN -------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=39578.63..39660.76 rows=821 width=41) (actual time=87805.47..101251.35 rows=144 loops=1) -> Group (cost=39578.63..39640.23 rows=8213 width=41) (actual time=87779.56..96824.56 rows=462198 loops=1) -> Sort (cost=39578.63..39599.17 rows=8213 width=41) (actual time=87779.52..90781.48 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=38899.14..39044.62 rows=8213 width=41) (actual time=64073.98..72783.68 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=64.66..66.55 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=1.76..61.70 rows=352 loops=1) -> Sort (cost=38874.73..38895.27 rows=8213 width=16) (actual time=64009.26..66860.71 rows=462198 loops=1) Sort Key: ts.company_id -> Seq Scan on traffic_logs ts (cost=0.00..38340.72 rows=8213 width=16) (actual time=5.02..-645982.04 rows=462198 loops=1) Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 101277.17 msec (14 rows) ---------------------------(end of broadcast)--------------------------- TIP 7: don't forget to increase your free space map settings From pgsql-performance-owner@postgresql.org Mon Nov 10 16:40:59 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DEBC0D1B550; Mon, 10 Nov 2003 20:40:57 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 80284-04; Mon, 10 Nov 2003 16:40:27 -0400 (AST) Received: from fddlnint05.fds.com (fddlnint05.fds.com [208.15.91.52]) by svr1.postgresql.org (Postfix) with ESMTP id D1881D1B53B; Mon, 10 Nov 2003 16:40:25 -0400 (AST) In-Reply-To: Subject: Re: *very* slow query to summarize data for a month ... To: pgsql-performance@postgresql.org, pgsql-performance-owner@postgresql.org, "Marc G. Fournier" X-Mailer: Lotus Notes Release 6.0 September 26, 2002 Message-ID: From: "Patrick Hatcher" Date: Mon, 10 Nov 2003 12:32:12 -0800 X-MIMETrack: Serialize by Router on FDDLNINT05/FSG/SVR/FDD(Release 5.0.4 |June 8, 2000) at 11/10/2003 03:34:52 PM MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/29 X-Sequence-Number: 4589 here's the URL: http://techdocs.postgresql.org/techdocs/pgsqladventuresep2.php Patrick Hatcher Macys.Com Legacy Integration Developer 415-422-1610 office HatcherPT - AIM Patrick Hatcher/MCOM/FDD To 11/10/2003 12:31 PM "Marc G. Fournier" @FDS-NOTES cc pgsql-performance@postgresql.org, pgsql-performance-owner@postgresql.o rg Subject Re: [PERFORM] *very* slow query to summarize data for a month ... (Document link: Patrick Hatcher) Do you have an index on ts.bytes? Josh had suggested this and after I put it on my summed fields, I saw a speed increase. I can't remember the article was that Josh had written about index usage, but maybe he'll chime in and supply the URL for his article. hth Patrick Hatcher "Marc G. Fournier" To Sent by: pgsql-performance@postgresql.org pgsql-performance-o cc wner@postgresql.org Subject [PERFORM] *very* slow query to 11/10/2003 12:18 PM summarize data for a month ... Table structure is simple: CREATE TABLE traffic_logs ( company_id bigint, ip_id bigint, port integer, bytes bigint, runtime timestamp without time zone ); runtime is 'day of month' ... I need to summarize the month, per company, with a query as: explain analyze SELECT ts.company_id, company_name, SUM(ts.bytes) AS total_traffic FROM company c, traffic_logs ts WHERE c.company_id = ts.company_id AND month_trunc(ts.runtime) = '2003-10-01' GROUP BY company_name,ts.company_id; and the explain looks like: QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=32000.94..32083.07 rows=821 width=41) (actual time=32983.36..47586.17 rows=144 loops=1) -> Group (cost=32000.94..32062.54 rows=8213 width=41) (actual time=32957.40..42817.88 rows=462198 loops=1) -> Sort (cost=32000.94..32021.47 rows=8213 width=41) (actual time=32957.38..36261.31 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=31321.45..31466.92 rows=8213 width=41) (actual time=13983.07..22642.14 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=5.52..7.40 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=0.02..2.78 rows=352 loops=1) -> Sort (cost=31297.04..31317.57 rows=8213 width=16) (actual time=13977.49..16794.41 rows=462198 loops=1) Sort Key: ts.company_id -> Index Scan using tl_month on traffic_logs ts (cost=0.00..30763.02 rows=8213 width=16) (actual time=0.29..5562.25 rows=462198 loops=1) Index Cond: (month_trunc(runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 47587.82 msec (14 rows) the problem is that we're only taking a few months worth of data, so I don't think there is much of a way of 'improve performance' on this, but figured I'd ask quickly before I do something rash ... Note that without the month_trunc() index, the Total runtime more then doubles: QUERY PLAN -------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=39578.63..39660.76 rows=821 width=41) (actual time=87805.47..101251.35 rows=144 loops=1) -> Group (cost=39578.63..39640.23 rows=8213 width=41) (actual time=87779.56..96824.56 rows=462198 loops=1) -> Sort (cost=39578.63..39599.17 rows=8213 width=41) (actual time=87779.52..90781.48 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=38899.14..39044.62 rows=8213 width=41) (actual time=64073.98..72783.68 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=64.66..66.55 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=1.76..61.70 rows=352 loops=1) -> Sort (cost=38874.73..38895.27 rows=8213 width=16) (actual time=64009.26..66860.71 rows=462198 loops=1) Sort Key: ts.company_id -> Seq Scan on traffic_logs ts (cost=0.00..38340.72 rows=8213 width=16) (actual time=5.02..-645982.04 rows=462198 loops=1) Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 101277.17 msec (14 rows) ---------------------------(end of broadcast)--------------------------- TIP 7: don't forget to increase your free space map settings From pgsql-performance-owner@postgresql.org Mon Nov 10 17:02:08 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3CB48D1B502; Mon, 10 Nov 2003 21:02:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 82717-05; Mon, 10 Nov 2003 17:01:36 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 7F794D1B4E1; Mon, 10 Nov 2003 17:01:34 -0400 (AST) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3896758; Mon, 10 Nov 2003 13:02:18 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: "Marc G. Fournier" , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... Date: Mon, 10 Nov 2003 12:59:03 -0800 User-Agent: KMail/1.4.3 References: <20031110160144.C727@ganymede.hub.org> In-Reply-To: <20031110160144.C727@ganymede.hub.org> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200311101259.03389.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/30 X-Sequence-Number: 4590 Marc, I'd say your machine is very low on available RAM, particularly sort_mem.= =20=20=20 The steps which are taking a long time are: =20 > Aggregate (cost=3D32000.94..32083.07 rows=3D821 width=3D41) (actual=20 time=3D32983.36..47586.17 rows=3D144 loops=3D1) > -> Group (cost=3D32000.94..32062.54 rows=3D8213 width=3D41) (actual= =20 time=3D32957.40..42817.88 rows=3D462198 loops=3D1) and: > -> Merge Join (cost=3D31321.45..31466.92 rows=3D8213 wid= th=3D41)=20 (actual time=3D13983.07..22642.14 rows=3D462198 loops=3D1) > Merge Cond: ("outer".company_id =3D "inner".company_= id) > -> Sort (cost=3D24.41..25.29 rows=3D352 width=3D25= ) (actual=20 time=3D5.52..7.40 rows=3D348 loops=3D1) There are also *large* delays between steps. Either your I/O is saturate= d,=20 or you haven't run a VACUUM FULL ANALYZE in a while (which would also expla= in=20 the estimates being off). --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Mon Nov 10 18:51:42 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EF0A4D1B532; Mon, 10 Nov 2003 22:51:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 02423-05; Mon, 10 Nov 2003 18:51:12 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id D4DE1D1B4E4; Mon, 10 Nov 2003 18:51:09 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 512951EB9; Mon, 10 Nov 2003 17:51:11 -0500 (EST) To: "Patrick Hatcher" Cc: scrappy@postgresql.org, pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... From: Neil Conway In-Reply-To: (Patrick Hatcher's message of "Mon, 10 Nov 2003 12:31:08 -0800") References: Date: Mon, 10 Nov 2003 17:51:10 -0500 Message-ID: <87u15bx2g1.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/31 X-Sequence-Number: 4591 "Patrick Hatcher" writes: > Do you have an index on ts.bytes? Josh had suggested this and after I put > it on my summed fields, I saw a speed increase. What's the reasoning behind this? ISTM that sum() should never use an index, nor would it benefit from using one. -Neil From pgsql-performance-owner@postgresql.org Mon Nov 10 19:16:16 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id ED640D1B4E4; Mon, 10 Nov 2003 23:16:14 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 02416-10; Mon, 10 Nov 2003 19:15:45 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id C86AAD1B50B; Mon, 10 Nov 2003 19:15:42 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 98E6E1E37; Mon, 10 Nov 2003 18:15:44 -0500 (EST) To: "Marc G. Fournier" Cc: pgsql-performance@postgresql.org, Tom Lane Subject: Re: *very* slow query to summarize data for a month ... From: Neil Conway In-Reply-To: <20031110160144.C727@ganymede.hub.org> (Marc G. Fournier's message of "Mon, 10 Nov 2003 16:18:52 -0400 (AST)") References: <20031110160144.C727@ganymede.hub.org> Date: Mon, 10 Nov 2003 18:15:41 -0500 Message-ID: <87ptfzx1b6.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/32 X-Sequence-Number: 4592 "Marc G. Fournier" writes: > -> Index Scan using tl_month on traffic_logs ts (cost=0.00..30763.02 rows=8213 width=16) (actual time=0.29..5562.25 rows=462198 loops=1) > Index Cond: (month_trunc(runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Interesting that we get the row count estimate for this index scan so wrong -- I believe this is the root of the problem. Hmmm... I would guess that the optimizer stats we have for estimating the selectivity of a functional index is pretty primitive, but I haven't looked into it at all. Tom might be able to shed some light... [ In the second EXPLAIN ANALYZE, ... ] > -> Seq Scan on traffic_logs ts (cost=0.00..38340.72 rows=8213 width=16) (actual time=5.02..-645982.04 rows=462198 loops=1) > Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Uh, what? The "actual time" seems to have finished far before it has begun :-) Is this just a typo, or does the actual output include a negative number? -Neil From pgsql-performance-owner@postgresql.org Mon Nov 10 19:43:18 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8BE8AD1B527; Mon, 10 Nov 2003 23:43:14 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 06338-09; Mon, 10 Nov 2003 19:42:46 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id F369FD1B504; Mon, 10 Nov 2003 19:42:06 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAANg919016201; Mon, 10 Nov 2003 18:42:09 -0500 (EST) To: Neil Conway Cc: "Marc G. Fournier" , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... In-reply-to: <87ptfzx1b6.fsf@mailbox.samurai.com> References: <20031110160144.C727@ganymede.hub.org> <87ptfzx1b6.fsf@mailbox.samurai.com> Comments: In-reply-to Neil Conway message dated "Mon, 10 Nov 2003 18:15:41 -0500" Date: Mon, 10 Nov 2003 18:42:09 -0500 Message-ID: <16200.1068507729@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/33 X-Sequence-Number: 4593 Neil Conway writes: > Interesting that we get the row count estimate for this index scan so > wrong -- I believe this is the root of the problem. Hmmm... I would > guess that the optimizer stats we have for estimating the selectivity > of a functional index is pretty primitive, but I haven't looked into > it at all. Tom might be able to shed some light... Try "none at all". I have speculated in the past that it would be worth gathering statistics about the contents of functional indexes, but it's still on the to-do-someday list. >> -> Seq Scan on traffic_logs ts (cost=0.00..38340.72 rows=8213 width=16) (actual time=5.02..-645982.04 rows=462198 loops=1) > Uh, what? That is bizarre, all right. Is it reproducible? regards, tom lane From pgsql-performance-owner@postgresql.org Mon Nov 10 20:22:37 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D5B88D1B522; Tue, 11 Nov 2003 00:22:35 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 15417-05; Mon, 10 Nov 2003 20:22:07 -0400 (AST) Received: from ganymede.hub.org (u46n208.hfx.eastlink.ca [24.222.46.208]) by svr1.postgresql.org (Postfix) with ESMTP id 20FE1D1B4F2; Mon, 10 Nov 2003 20:22:04 -0400 (AST) Received: by ganymede.hub.org (Postfix, from userid 1000) id 0789836F68; Mon, 10 Nov 2003 20:19:57 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by ganymede.hub.org (Postfix) with ESMTP id EDC2236F64; Mon, 10 Nov 2003 20:19:56 -0400 (AST) Date: Mon, 10 Nov 2003 20:19:56 -0400 (AST) From: "Marc G. Fournier" X-X-Sender: scrappy@ganymede.hub.org To: Neil Conway Cc: "Marc G. Fournier" , pgsql-performance@postgresql.org, Tom Lane Subject: Re: *very* slow query to summarize data for a month ... In-Reply-To: <87ptfzx1b6.fsf@mailbox.samurai.com> Message-ID: <20031110201945.X727@ganymede.hub.org> References: <20031110160144.C727@ganymede.hub.org> <87ptfzx1b6.fsf@mailbox.samurai.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/34 X-Sequence-Number: 4594 On Mon, 10 Nov 2003, Neil Conway wrote: > "Marc G. Fournier" writes: > > -> Index Scan using tl_month on traffic_logs ts (cost=0.00..30763.02 rows=8213 width=16) (actual time=0.29..5562.25 rows=462198 loops=1) > > Index Cond: (month_trunc(runtime) = '2003-10-01 00:00:00'::timestamp without time zone) > > Interesting that we get the row count estimate for this index scan so > wrong -- I believe this is the root of the problem. Hmmm... I would > guess that the optimizer stats we have for estimating the selectivity > of a functional index is pretty primitive, but I haven't looked into > it at all. Tom might be able to shed some light... > > [ In the second EXPLAIN ANALYZE, ... ] > > > -> Seq Scan on traffic_logs ts (cost=0.00..38340.72 rows=8213 width=16) (actual time=5.02..-645982.04 rows=462198 loops=1) > > Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) > > Uh, what? The "actual time" seems to have finished far before it has > begun :-) Is this just a typo, or does the actual output include a > negative number? This was purely a cut-n-paste ... From pgsql-performance-owner@postgresql.org Mon Nov 10 20:30:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5696FD1B537; Tue, 11 Nov 2003 00:30:45 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 14342-08; Mon, 10 Nov 2003 20:30:17 -0400 (AST) Received: from ganymede.hub.org (u46n208.hfx.eastlink.ca [24.222.46.208]) by svr1.postgresql.org (Postfix) with ESMTP id 6DAE4D1B527; Mon, 10 Nov 2003 20:30:14 -0400 (AST) Received: by ganymede.hub.org (Postfix, from userid 1000) id 4A50937009; Mon, 10 Nov 2003 20:28:07 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by ganymede.hub.org (Postfix) with ESMTP id 3CD1A36F64; Mon, 10 Nov 2003 20:28:07 -0400 (AST) Date: Mon, 10 Nov 2003 20:28:07 -0400 (AST) From: "Marc G. Fournier" X-X-Sender: scrappy@ganymede.hub.org To: Tom Lane Cc: Neil Conway , "Marc G. Fournier" , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... In-Reply-To: <16200.1068507729@sss.pgh.pa.us> Message-ID: <20031110202726.J727@ganymede.hub.org> References: <20031110160144.C727@ganymede.hub.org> <87ptfzx1b6.fsf@mailbox.samurai.com> <16200.1068507729@sss.pgh.pa.us> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/35 X-Sequence-Number: 4595 On Mon, 10 Nov 2003, Tom Lane wrote: > Neil Conway writes: > > Interesting that we get the row count estimate for this index scan so > > wrong -- I believe this is the root of the problem. Hmmm... I would > > guess that the optimizer stats we have for estimating the selectivity > > of a functional index is pretty primitive, but I haven't looked into > > it at all. Tom might be able to shed some light... > > Try "none at all". I have speculated in the past that it would be worth > gathering statistics about the contents of functional indexes, but it's > still on the to-do-someday list. > > >> -> Seq Scan on traffic_logs ts (cost=0.00..38340.72 rows=8213 width=16) (actual time=5.02..-645982.04 rows=462198 loops=1) > > > Uh, what? > > That is bizarre, all right. Is it reproducible? Nope, and a subsequent run shows better results too: QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=39674.38..39756.70 rows=823 width=41) (actual time=35573.27..49953.47 rows=144 loops=1) -> Group (cost=39674.38..39736.12 rows=8232 width=41) (actual time=35547.27..45479.27 rows=462198 loops=1) -> Sort (cost=39674.38..39694.96 rows=8232 width=41) (actual time=35547.23..39167.90 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=38993.22..39139.02 rows=8232 width=41) (actual time=16658.23..25559.08 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=5.51..7.38 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=0.02..2.80 rows=352 loops=1) -> Sort (cost=38968.82..38989.40 rows=8232 width=16) (actual time=16652.66..19785.83 rows=462198 loops=1) Sort Key: ts.company_id -> Seq Scan on traffic_logs ts (cost=0.00..38433.46 rows=8232 width=16) (actual time=0.11..8794.43 rows=462198 loops=1) Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 49955.22 msec From pgsql-performance-owner@postgresql.org Mon Nov 10 20:52:40 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1993FD1B532; Tue, 11 Nov 2003 00:52:37 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 23471-03; Mon, 10 Nov 2003 20:52:08 -0400 (AST) Received: from ganymede.hub.org (u46n208.hfx.eastlink.ca [24.222.46.208]) by svr1.postgresql.org (Postfix) with ESMTP id 6DE17D1B53F; Mon, 10 Nov 2003 20:52:04 -0400 (AST) Received: by ganymede.hub.org (Postfix, from userid 1000) id AFBEC36BEF; Mon, 10 Nov 2003 20:49:57 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by ganymede.hub.org (Postfix) with ESMTP id A2A753661D; Mon, 10 Nov 2003 20:49:57 -0400 (AST) Date: Mon, 10 Nov 2003 20:49:57 -0400 (AST) From: "Marc G. Fournier" X-X-Sender: scrappy@ganymede.hub.org To: Josh Berkus Cc: "Marc G. Fournier" , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... In-Reply-To: <200311101259.03389.josh@agliodbs.com> Message-ID: <20031110203259.Y727@ganymede.hub.org> References: <20031110160144.C727@ganymede.hub.org> <200311101259.03389.josh@agliodbs.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/36 X-Sequence-Number: 4596 On Mon, 10 Nov 2003, Josh Berkus wrote: > Marc, > > I'd say your machine is very low on available RAM, particularly sort_mem. > The steps which are taking a long time are: Here's the server: last pid: 42651; load averages: 1.52, 0.96, 0.88 up 28+07:43:33 20:35:44 307 processes: 2 running, 304 sleeping, 1 zombie CPU states: 18.0% user, 0.0% nice, 29.1% system, 0.6% interrupt, 52.3% idle Mem: 1203M Active, 1839M Inact, 709M Wired, 206M Cache, 199M Buf, 5608K Free Swap: 8192M Total, 1804K Used, 8190M Free > > > Aggregate (cost=32000.94..32083.07 rows=821 width=41) (actual > time=32983.36..47586.17 rows=144 loops=1) > > -> Group (cost=32000.94..32062.54 rows=8213 width=41) (actual > time=32957.40..42817.88 rows=462198 loops=1) > > and: > > > -> Merge Join (cost=31321.45..31466.92 rows=8213 width=41) > (actual time=13983.07..22642.14 rows=462198 loops=1) > > Merge Cond: ("outer".company_id = "inner".company_id) > > -> Sort (cost=24.41..25.29 rows=352 width=25) (actual > time=5.52..7.40 rows=348 loops=1) > > There are also *large* delays between steps. Either your I/O is saturated, > or you haven't run a VACUUM FULL ANALYZE in a while (which would also explain > the estimates being off). thought about that before I started the thread, and ran it just in case ... just restarted the server with sort_mem set to 10M, and didn't help much on the Aggregate, or MergeJoin ... : QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=39674.38..39756.70 rows=823 width=41) (actual time=33066.25..54021.50 rows=144 loops=1) -> Group (cost=39674.38..39736.12 rows=8232 width=41) (actual time=33040.25..47005.57 rows=462198 loops=1) -> Sort (cost=39674.38..39694.96 rows=8232 width=41) (actual time=33040.22..37875.97 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=38993.22..39139.02 rows=8232 width=41) (actual time=14428.17..23568.80 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=5.80..7.66 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=0.08..3.06 rows=352 loops=1) -> Sort (cost=38968.82..38989.40 rows=8232 width=16) (actual time=14422.27..17429.34 rows=462198 loops=1) Sort Key: ts.company_id -> Seq Scan on traffic_logs ts (cost=0.00..38433.46 rows=8232 width=16) (actual time=0.15..8119.72 rows=462198 loops=1) Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 54034.44 msec (14 rows) the problem is that the results we are comparing with right now is the one that had the - time on it :( Just restarted the server with default sort_mem, and here is the query with that: QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=39691.27..39773.61 rows=823 width=41) (actual time=35077.18..50424.74 rows=144 loops=1) -> Group (cost=39691.27..39753.03 rows=8234 width=41) (actual time=35051.29..-650049.84 rows=462198 loops=1) -> Sort (cost=39691.27..39711.86 rows=8234 width=41) (actual time=35051.26..38847.40 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=39009.92..39155.76 rows=8234 width=41) (actual time=16155.37..25439.42 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=5.85..7.71 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=0.10..3.07 rows=352 loops=1) -> Sort (cost=38985.51..39006.10 rows=8234 width=16) (actual time=16149.46..19437.47 rows=462198 loops=1) Sort Key: ts.company_id -> Seq Scan on traffic_logs ts (cost=0.00..38450.00 rows=8234 width=16) (actual time=0.16..8869.37 rows=462198 loops=1) Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 50426.80 msec (14 rows) And, just on a whim, here it is set to 100M: QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=39691.27..39773.61 rows=823 width=41) (actual time=25888.20..38909.88 rows=144 loops=1) -> Group (cost=39691.27..39753.03 rows=8234 width=41) (actual time=25862.81..34591.76 rows=462198 loops=1) -> Sort (cost=39691.27..39711.86 rows=8234 width=41) (actual time=25862.77..723885.95 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=39009.92..39155.76 rows=8234 width=41) (actual time=12471.23..21855.08 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=5.87..7.74 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=0.11..3.14 rows=352 loops=1) -> Sort (cost=38985.51..39006.10 rows=8234 width=16) (actual time=12465.29..14941.24 rows=462198 loops=1) Sort Key: ts.company_id -> Seq Scan on traffic_logs ts (cost=0.00..38450.00 rows=8234 width=16) (actual time=0.18..9106.16 rows=462198 loops=1) Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 39077.75 msec (14 rows) So, it does give a noticeable improvement the higher the sort_mem ... And, @ 100M for sort_mem and using the month_trunc index: QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=32089.29..32171.63 rows=823 width=41) (actual time=30822.51..57202.44 rows=144 loops=1) -> Group (cost=32089.29..32151.04 rows=8234 width=41) (actual time=30784.24..743396.18 rows=462198 loops=1) -> Sort (cost=32089.29..32109.87 rows=8234 width=41) (actual time=30784.21..36212.96 rows=462198 loops=1) Sort Key: c.company_name, ts.company_id -> Merge Join (cost=31407.94..31553.77 rows=8234 width=41) (actual time=11384.79..24918.56 rows=462198 loops=1) Merge Cond: ("outer".company_id = "inner".company_id) -> Sort (cost=24.41..25.29 rows=352 width=25) (actual time=5.92..9.55 rows=348 loops=1) Sort Key: c.company_id -> Seq Scan on company c (cost=0.00..9.52 rows=352 width=25) (actual time=0.08..3.21 rows=352 loops=1) -> Sort (cost=31383.53..31404.12 rows=8234 width=16) (actual time=11378.81..15211.07 rows=462198 loops=1) Sort Key: ts.company_id -> Index Scan using tl_month on traffic_logs ts (cost=0.00..30848.02 rows=8234 width=16) (actual time=0.46..7055.75 rows=462198 loops=1) Index Cond: (month_trunc(runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 57401.72 msec (14 rows) From pgsql-performance-owner@postgresql.org Tue Nov 11 02:50:43 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2ABAED1B502; Tue, 11 Nov 2003 06:50:41 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 84838-06; Tue, 11 Nov 2003 02:50:11 -0400 (AST) Received: from zigo.dhs.org (as2-4-3.an.g.bonet.se [194.236.34.191]) by svr1.postgresql.org (Postfix) with ESMTP id 3E602D1B4E1; Tue, 11 Nov 2003 02:50:09 -0400 (AST) Received: from zigo.dhs.org (zigo [127.0.0.1]) by zigo.dhs.org (8.12.8/8.12.8) with ESMTP id hAB6o8L2029244; Tue, 11 Nov 2003 07:50:08 +0100 Received: from localhost (db@localhost) by zigo.dhs.org (8.12.8/8.12.8/Submit) with ESMTP id hAB6o7FY029239; Tue, 11 Nov 2003 07:50:08 +0100 Date: Tue, 11 Nov 2003 07:50:07 +0100 (CET) From: Dennis Bjorklund To: "Marc G. Fournier" Cc: pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... In-Reply-To: <20031110160144.C727@ganymede.hub.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/37 X-Sequence-Number: 4597 On Mon, 10 Nov 2003, Marc G. Fournier wrote: > > explain analyze SELECT ts.company_id, company_name, SUM(ts.bytes) AS total_traffic > FROM company c, traffic_logs ts > WHERE c.company_id = ts.company_id > AND month_trunc(ts.runtime) = '2003-10-01' > GROUP BY company_name,ts.company_id; What if you do ts.runtime >= '2003-10-01' AND ts.runtime < '2003-11-01' and add an index like (runtime, company_name, company_id)? -- /Dennis From pgsql-performance-owner@postgresql.org Tue Nov 11 13:47:56 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 79EEBD1B564; Tue, 11 Nov 2003 17:47:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 88841-08; Tue, 11 Nov 2003 13:47:23 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 23559D1B51E; Tue, 11 Nov 2003 13:47:22 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id C15FC36A82; Tue, 11 Nov 2003 12:47:06 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1AJcba-0007BP-00; Tue, 11 Nov 2003 12:47:06 -0500 To: Dennis Bjorklund Cc: "Marc G. Fournier" , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... References: In-Reply-To: From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 11 Nov 2003 12:47:06 -0500 Message-ID: <878ymmkdb9.fsf@stark.dyndns.tv> Lines: 41 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/40 X-Sequence-Number: 4600 Dennis Bjorklund writes: > On Mon, 10 Nov 2003, Marc G. Fournier wrote: > > > > > explain analyze SELECT ts.company_id, company_name, SUM(ts.bytes) AS total_traffic > > FROM company c, traffic_logs ts > > WHERE c.company_id = ts.company_id > > AND month_trunc(ts.runtime) = '2003-10-01' > > GROUP BY company_name,ts.company_id; So depending on how much work you're willing to do there are some more dramatic speedups you could get: Use partial indexes like this (you'll need one for every month): create index i on traffic_log (company_id) where month_trunc(runtime) = '2003-10-01' then group by company_id only so it can use the index: select * from company join ( select company_id, sum(bytes) as total_traffic from traffic_log where month_trunc(runtime) = '2003-10-01' group by company_id ) as x using (company_id) order by company_name Actually you might be able to get the same effect using function indexes like: create index i on traffic_log (month_trunc(runtime), company_id) -- greg From pgsql-performance-owner@postgresql.org Tue Nov 11 14:19:02 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E4763D1B53F for ; Tue, 11 Nov 2003 18:18:59 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 00214-05 for ; Tue, 11 Nov 2003 14:18:29 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 5CD7FD1B53B for ; Tue, 11 Nov 2003 14:18:21 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hABIILNu086094 for ; Tue, 11 Nov 2003 18:18:21 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hABICUNC085108 for pgsql-performance@postgresql.org; Tue, 11 Nov 2003 18:12:30 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Suggestions for benchmarking 7.4RC2 against 7.3 Date: Tue, 11 Nov 2003 13:08:08 -0500 Organization: cbbrowne Computing Inc Lines: 16 Message-ID: References: <3FB247FE.4080808@trade-india.com> <3FB24C15.1020907@trade-india.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:VfvMkXcJ1gw579SqoonCd8FvsIw= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/42 X-Sequence-Number: 4602 After a long battle with technology,mallah@trade-india.com (Rajesh Kumar Mallah), an earthling, wrote: > the error mentioned in first email has been overcome > by running osdb on the same machine hosting the DB server. Yes, it seems unrealistic to try to run the "client" on a separate host from the database. I got the osdb benchmark running last week, and had to separate client from server. I had to jump through a fair number of hoops including copying data files over to the server. The benchmark software needs a bit more work... -- let name="cbbrowne" and tld="cbbrowne.com" in String.concat "@" [name;tld];; http://cbbrowne.com/info/lsf.html Nobody can fix the economy. Nobody can be trusted with their finger on the button. Nobody's perfect. VOTE FOR NOBODY. From pgsql-performance-owner@postgresql.org Tue Nov 11 14:16:48 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 13C71D1B544 for ; Tue, 11 Nov 2003 18:16:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 00243-01 for ; Tue, 11 Nov 2003 14:16:16 -0400 (AST) Received: from ganymede.hub.org (u46n208.hfx.eastlink.ca [24.222.46.208]) by svr1.postgresql.org (Postfix) with ESMTP id E9DDED1B550 for ; Tue, 11 Nov 2003 14:16:08 -0400 (AST) Received: by ganymede.hub.org (Postfix, from userid 1000) id 4F6D837603; Tue, 11 Nov 2003 14:14:01 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by ganymede.hub.org (Postfix) with ESMTP id 4E99D37512; Tue, 11 Nov 2003 14:14:01 -0400 (AST) Date: Tue, 11 Nov 2003 14:14:01 -0400 (AST) From: "Marc G. Fournier" X-X-Sender: scrappy@ganymede.hub.org To: Greg Stark Cc: Dennis Bjorklund , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... In-Reply-To: <878ymmkdb9.fsf@stark.dyndns.tv> Message-ID: <20031111141328.X56037@ganymede.hub.org> References: <878ymmkdb9.fsf@stark.dyndns.tv> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/41 X-Sequence-Number: 4601 On Tue, 11 Nov 2003, Greg Stark wrote: > Actually you might be able to get the same effect using function indexes > like: > > create index i on traffic_log (month_trunc(runtime), company_id) had actually thought of that one ... is it something that is only available in v7.4? ams=# create index i on traffic_logs ( month_trunc(runtime), company_id ); ERROR: parser: parse error at or near "," at character 54 From pgsql-performance-owner@postgresql.org Tue Nov 11 14:26:11 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id BDADED1B53B for ; Tue, 11 Nov 2003 18:26:10 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 99727-07 for ; Tue, 11 Nov 2003 14:25:40 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id CE3E2D1B534 for ; Tue, 11 Nov 2003 14:25:38 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3901393; Tue, 11 Nov 2003 10:26:24 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: Christopher Browne , pgsql-performance@postgresql.org Subject: Re: Suggestions for benchmarking 7.4RC2 against 7.3 Date: Tue, 11 Nov 2003 10:25:28 -0800 User-Agent: KMail/1.4.3 References: <3FB247FE.4080808@trade-india.com> <3FB24C15.1020907@trade-india.com> In-Reply-To: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311111025.28806.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/43 X-Sequence-Number: 4603 Rajesh, Chris, > I got the osdb benchmark running last week, and had to separate client > from server. I had to jump through a fair number of hoops including > copying data files over to the server. The benchmark software needs a > bit more work... What about the OSDL's TPC-derivative benchmarks? That's a much more respected database test, and probably less buggy than OSDB. -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Nov 11 14:31:23 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B6C80D1B53F; Tue, 11 Nov 2003 18:31:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 00243-08; Tue, 11 Nov 2003 14:30:51 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id B1486D1B53B; Tue, 11 Nov 2003 14:30:49 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3901415; Tue, 11 Nov 2003 10:31:35 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: "Marc G. Fournier" , Greg Stark Subject: Re: *very* slow query to summarize data for a month ... Date: Tue, 11 Nov 2003 10:30:39 -0800 User-Agent: KMail/1.4.3 Cc: Dennis Bjorklund , pgsql-performance@postgresql.org References: <878ymmkdb9.fsf@stark.dyndns.tv> <20031111141328.X56037@ganymede.hub.org> In-Reply-To: <20031111141328.X56037@ganymede.hub.org> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311111030.39496.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/44 X-Sequence-Number: 4604 marc, > had actually thought of that one ... is it something that is only > available in v7.4? Yes. New feature. -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Nov 11 14:40:44 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4E220D1B54C; Tue, 11 Nov 2003 18:40:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 00809-07; Tue, 11 Nov 2003 14:40:13 -0400 (AST) Received: from ganymede.hub.org (u46n208.hfx.eastlink.ca [24.222.46.208]) by svr1.postgresql.org (Postfix) with ESMTP id 3E65ED1B521; Tue, 11 Nov 2003 14:40:12 -0400 (AST) Received: by ganymede.hub.org (Postfix, from userid 1000) id BD5B236BD7; Tue, 11 Nov 2003 14:38:04 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by ganymede.hub.org (Postfix) with ESMTP id BA82436B2E; Tue, 11 Nov 2003 14:38:04 -0400 (AST) Date: Tue, 11 Nov 2003 14:38:04 -0400 (AST) From: "Marc G. Fournier" X-X-Sender: scrappy@ganymede.hub.org To: Dennis Bjorklund Cc: "Marc G. Fournier" , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... In-Reply-To: Message-ID: <20031111134925.L56037@ganymede.hub.org> References: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/45 X-Sequence-Number: 4605 On Tue, 11 Nov 2003, Dennis Bjorklund wrote: > On Mon, 10 Nov 2003, Marc G. Fournier wrote: > > > > > explain analyze SELECT ts.company_id, company_name, SUM(ts.bytes) AS total_traffic > > FROM company c, traffic_logs ts > > WHERE c.company_id = ts.company_id > > AND month_trunc(ts.runtime) = '2003-10-01' > > GROUP BY company_name,ts.company_id; > > What if you do > > ts.runtime >= '2003-10-01' AND ts.runtime < '2003-11-01' > > and add an index like (runtime, company_name, company_id)? Good thought, but even simplifying it to the *lowest* query possible, with no table joins, is painfully slow: explain analyze SELECT ts.company_id, SUM(ts.bytes) AS total_traffic FROM traffic_logs ts WHERE month_trunc(ts.runtime) = '2003-10-01' GROUP BY ts.company_id; QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=31630.84..31693.05 rows=829 width=16) (actual time=14862.71..26552.39 rows=144 loops=1) -> Group (cost=31630.84..31672.31 rows=8295 width=16) (actual time=9634.28..20967.07 rows=462198 loops=1) -> Sort (cost=31630.84..31651.57 rows=8295 width=16) (actual time=9634.24..12838.73 rows=462198 loops=1) Sort Key: company_id -> Index Scan using tl_month on traffic_logs ts (cost=0.00..31090.93 rows=8295 width=16) (actual time=0.26..6043.35 rows=462198 loops=1) Index Cond: (month_trunc(runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 26659.35 msec (7 rows) -OR- explain analyze SELECT ts.company_id, SUM(ts.bytes) AS total_traffic FROM traffic_logs ts WHERE ts.runtime >= '2003-10-01' AND ts.runtime < '2003-11-01' GROUP BY ts.company_id; QUERY PLAN ------------------------------------------------------------------------------------------------------------------------------------------------------------------ Aggregate (cost=81044.53..84424.21 rows=45062 width=16) (actual time=13307.52..29274.66 rows=144 loops=1) -> Group (cost=81044.53..83297.65 rows=450625 width=16) (actual time=10809.02..-673265.13 rows=462198 loops=1) -> Sort (cost=81044.53..82171.09 rows=450625 width=16) (actual time=10808.99..14069.79 rows=462198 loops=1) Sort Key: company_id -> Seq Scan on traffic_logs ts (cost=0.00..38727.35 rows=450625 width=16) (actual time=0.07..6801.92 rows=462198 loops=1) Filter: ((runtime >= '2003-10-01 00:00:00'::timestamp without time zone) AND (runtime < '2003-11-01 00:00:00'::timestamp without time zone)) Total runtime: 29385.97 msec (7 rows) Just as a side note, just doing a straight scan for the records, with no SUM()/GROUP BY involved, with the month_trunc() index is still >8k msec: QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------------------- Index Scan using tl_month on traffic_logs ts (cost=0.00..31096.36 rows=8297 width=16) (actual time=0.96..5432.93 rows=462198 loops=1) Index Cond: (month_trunc(runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 8092.88 msec (3 rows) and without the index, >15k msec: QUERY PLAN ------------------------------------------------------------------------------------------------------------------------ Seq Scan on traffic_logs ts (cost=0.00..38719.55 rows=8297 width=16) (actual time=0.11..11354.45 rows=462198 loops=1) Filter: (date_trunc('month'::text, runtime) = '2003-10-01 00:00:00'::timestamp without time zone) Total runtime: 15353.57 msec (3 rows) so the GROUP BY is affecting the overall, but even without it, its still taking a helluva long time ... I'm going to modify my load script so that it dumps monthly totals to traffic_logs, and 'details' to a schema.traffic_logs table ... I don't need the 'per day totals' at the top level at all, only speed ... the 'per day totals' are only required at the 'per client' level, and by moving the 'per day' into a client schema will shrink the table significantly ... If it wasn't for trying to pull in that 'whole month' summary, it would be fine :( From pgsql-performance-owner@postgresql.org Tue Nov 11 15:51:54 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 064C5D1B50D; Tue, 11 Nov 2003 19:51:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 11750-10; Tue, 11 Nov 2003 15:51:23 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 29BF3D1B4FE; Tue, 11 Nov 2003 15:51:22 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id 9826F36AD6; Tue, 11 Nov 2003 14:51:22 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1AJeXq-0000EA-00; Tue, 11 Nov 2003 14:51:22 -0500 To: "Marc G. Fournier" Cc: Greg Stark , Dennis Bjorklund , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... References: <878ymmkdb9.fsf@stark.dyndns.tv> <20031111141328.X56037@ganymede.hub.org> In-Reply-To: <20031111141328.X56037@ganymede.hub.org> From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 11 Nov 2003 14:51:22 -0500 Message-ID: <873ccuk7k5.fsf@stark.dyndns.tv> Lines: 21 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/46 X-Sequence-Number: 4606 "Marc G. Fournier" writes: > On Tue, 11 Nov 2003, Greg Stark wrote: > > > Actually you might be able to get the same effect using function indexes > > like: > > > > create index i on traffic_log (month_trunc(runtime), company_id) > > had actually thought of that one ... is it something that is only > available in v7.4? Hum, I thought you could do simple functional indexes like that in 7.3, but perhaps only single-column indexes. In any case, given your situation I would seriously consider putting a "month" integer column on your table anyways. Then your index would be a simple (month, company_id) index. -- greg From pgsql-performance-owner@postgresql.org Tue Nov 11 17:40:15 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 10010D1B535; Tue, 11 Nov 2003 21:40:14 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 32471-05; Tue, 11 Nov 2003 17:39:44 -0400 (AST) Received: from mail1.ihs.com (mail1.ihs.com [170.207.70.222]) by svr1.postgresql.org (Postfix) with ESMTP id ADF3FD1B4F6; Tue, 11 Nov 2003 17:39:41 -0400 (AST) Received: from css120.ihs.com (css120.ihs.com [170.207.105.120]) by mail1.ihs.com (8.12.10/8.12.10) with ESMTP id hABLdANK022676; Tue, 11 Nov 2003 14:39:11 -0700 (MST) Date: Tue, 11 Nov 2003 14:25:06 -0700 (MST) From: "scott.marlowe" To: Greg Stark Cc: "Marc G. Fournier" , Dennis Bjorklund , Subject: Re: *very* slow query to summarize data for a month ... In-Reply-To: <873ccuk7k5.fsf@stark.dyndns.tv> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-IHS-MailScanner: Found to be clean X-IHS-MailScanner-SpamCheck: X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/49 X-Sequence-Number: 4609 On 11 Nov 2003, Greg Stark wrote: > "Marc G. Fournier" writes: > > > On Tue, 11 Nov 2003, Greg Stark wrote: > > > > > Actually you might be able to get the same effect using function indexes > > > like: > > > > > > create index i on traffic_log (month_trunc(runtime), company_id) > > > > had actually thought of that one ... is it something that is only > > available in v7.4? > > Hum, I thought you could do simple functional indexes like that in 7.3, but > perhaps only single-column indexes. > > In any case, given your situation I would seriously consider putting a > "month" integer column on your table anyways. Then your index would be a > simple (month, company_id) index. In 7.3 and before, you had to use only column names as inputs, so you could cheat: alter table test add alp int; alter table test add omg int; update test set alp=0; update test set omg=13; and then create a functional index: create index test_xy on test (substr(info,alp,omg)); select * from test where substr(info,alp,omg)=='abcd'; From pgsql-performance-owner@postgresql.org Wed Nov 12 15:45:59 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B6B55D1B559 for ; Tue, 11 Nov 2003 22:27:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 36934-10 for ; Tue, 11 Nov 2003 18:27:19 -0400 (AST) Received: from shire.ontko.com (shire.ontko.com [199.164.165.1]) by svr1.postgresql.org (Postfix) with ESMTP id 0D9FED1B54E for ; Tue, 11 Nov 2003 18:27:00 -0400 (AST) Received: from nick (bilbo.ontko.com [199.164.165.101]) by shire.ontko.com (8.12.3/8.12.3/Debian-6.6) with SMTP id hABMQx5f032534 for ; Tue, 11 Nov 2003 17:27:02 -0500 Reply-To: From: "Nick Fankhauser - Doxpop" To: "Pgsql-Performance@Postgresql. Org" Subject: Seeking help with a query that take too long Date: Tue, 11 Nov 2003 17:26:48 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/76 X-Sequence-Number: 4636 Hi- I have a query that takes too long. I haven't been able to come up with any ideas for speeding it up, so I'm seeking some input from the list. I'm using version 7.3.2 I have three tables: case_data (1,947,386 rows) actor (3,385,669 rows) actor_case_assignment (8,668,650 rows) As the names imply, actor_case_assignment contains records that assign an actor to a case. Actors such as attorneys or judges may have many cases, while the average actor (we hope) only has one. What I'm trying to do is link these tables to get back a single row per actor that shows the actor's name, the number of cases that actor is assigned to, and if they only have one case, I want the public_id for that case. This means I have to do a group by to get the case count, but I'm then forced to use an aggregate function like max on the other fields. All of the fields ending in "_id" have unique indexes, and actor_full_name_uppercase is indexed. Here's the select: select actor.actor_id, max(actor.actor_full_name), max(case_data.case_public_id), max(case_data.case_id), count(case_data.case_id) as case_count from actor, actor_case_assignment, case_data where actor.actor_full_name_uppercase like upper('sanders%') and actor.actor_id = actor_case_assignment.actor_id and case_data.case_id = actor_case_assignment.case_id group by actor.actor_id order by max(actor.actor_full_name), case_count desc limit 1000; Here's the explain analyze: QUERY PLAN ---------------------------------------------------------------------------- ---------------------------------------------------------------------------- -------------------------------------------------- Limit (cost=2214.71..2214.72 rows=1 width=115) (actual time=120034.61..120035.67 rows=1000 loops=1) -> Sort (cost=2214.71..2214.72 rows=1 width=115) (actual time=120034.60..120034.98 rows=1001 loops=1) Sort Key: max((actor.actor_full_name)::text), count(case_data.case_id) -> Aggregate (cost=2214.67..2214.70 rows=1 width=115) (actual time=119962.80..120011.49 rows=3456 loops=1) -> Group (cost=2214.67..2214.68 rows=2 width=115) (actual time=119962.76..119987.04 rows=5879 loops=1) -> Sort (cost=2214.67..2214.68 rows=2 width=115) (actual time=119962.74..119965.09 rows=5879 loops=1) Sort Key: actor.actor_id -> Nested Loop (cost=0.00..2214.66 rows=2 width=115) (actual time=59.05..119929.71 rows=5879 loops=1) -> Nested Loop (cost=0.00..2205.26 rows=3 width=76) (actual time=51.46..66089.04 rows=5882 loops=1) -> Index Scan using actor_full_name_uppercase on actor (cost=0.00..6.01 rows=1 width=42) (actual time=37.62..677.44 rows=3501 loops=1) Index Cond: ((actor_full_name_uppercase >= 'SANDERS'::character varying) AND (actor_full_name_uppercase < 'SANDERT'::character varying)) Filter: (actor_full_name_uppercase ~~ 'SANDERS%'::text) -> Index Scan using actor_case_assignment_actor_id on actor_case_assignment (cost=0.00..2165.93 rows=2666 width=34) (actual time=16.37..18.67 rows=2 loops=3501) Index Cond: ("outer".actor_id = actor_case_assignment.actor_id) -> Index Scan using case_data_case_id on case_data (cost=0.00..3.66 rows=1 width=39) (actual time=9.14..9.15 rows=1 loops=5882) Index Cond: (case_data.case_id = "outer".case_id) Total runtime: 120038.60 msec (17 rows) Any ideas? Thanks! -Nick --------------------------------------------------------------------- Nick Fankhauser nickf@doxpop.com Phone 1.765.965.7363 Fax 1.765.962.9788 doxpop - Court records at your fingertips - http://www.doxpop.com/ From pgsql-performance-owner@postgresql.org Tue Nov 11 19:41:54 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CC937D1B531 for ; Tue, 11 Nov 2003 23:41:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 54023-10 for ; Tue, 11 Nov 2003 19:41:25 -0400 (AST) Received: from vt-pe2550-001.VANTAGE.vantage.com (unknown [64.80.203.242]) by svr1.postgresql.org (Postfix) with ESMTP id 1EB56D1B51E for ; Tue, 11 Nov 2003 19:41:22 -0400 (AST) X-MimeOLE: Produced By Microsoft Exchange V6.0.6249.0 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: Re: Server Configs Date: Tue, 11 Nov 2003 17:57:15 -0500 Message-ID: <2F2E24372F10744588A27DEECC85FE04023C416E@vt-pe2550-001.vantage.vantage.com> X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: [PERFORM] Server Configs Thread-Index: AcOjxEI2VXnAdXMxRXy5Wh4EYnKxpAE4LlDg From: "Anjan Dave" To: X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests= X-Spam-Level: X-Archive-Number: 200311/51 X-Sequence-Number: 4611 Dear Gurus, We are planning to add more db server hardware for the apps. The question is, what makes more sense regarding performance/scalability/price of the hardware... There are a couple of apps, currently on a dual-cpu Dell server. The usage of the apps is going to increase quite a lot, and considering the prices, we are looking at the following options: Option 1: =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Have each app on a separate db server (looking at 4 of these). The server being a PowerEdge 2650, Dual 2.8GHz/512KB XEONS, 2GB RAM, PERC-3 RAID-5, split back plane (2+3), and 5 x 36GB HDDs (10K RPM). Note: These servers are 1/3 the price of the Quad-cpu 6650 server. Option 2: =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Have two to three apps dbs hosted on a single server. The server being a PowerEdge 6650, 4 x 2GHz/1MB XEONS, 8GB RAM, PERC-3 RAID-5, split back plane (2+3), and 5 x 36GB HDDs (10K RPM). Note: This server is 3 times more the price of the option 1. Appreciate your guidance. Thanks, Anjan From pgsql-performance-owner@postgresql.org Tue Nov 11 19:33:35 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 155E5D1B51E for ; Tue, 11 Nov 2003 23:33:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 52424-08 for ; Tue, 11 Nov 2003 19:33:05 -0400 (AST) Received: from corp1.affinitysolutions.com (corp1.affinitysolutions.com [216.52.88.186]) by svr1.postgresql.org (Postfix) with ESMTP id 026ADD1B50C for ; Tue, 11 Nov 2003 19:33:03 -0400 (AST) Received: from tuxsrva.afsnyc-nt1.affinitysolutions.com (unknown [216.220.102.122]) by corp1.affinitysolutions.com (Postfix) with ESMTP id 06B9B11F705 for ; Tue, 11 Nov 2003 18:32:49 -0500 (EST) Subject: Value of Quad vs. Dual Processor machine From: Chris Field Reply-To: cfield@affinitysolutions.com To: pgsql-performance@postgresql.org Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-Pp9U7L6ZAviF2yFBgYtQ" Organization: Affinity Solutions Message-Id: <1068593567.2063.8.camel@tuxsrva> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.4 Date: Tue, 11 Nov 2003 18:32:47 -0500 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/50 X-Sequence-Number: 4610 --=-Pp9U7L6ZAviF2yFBgYtQ Content-Type: text/plain Content-Transfer-Encoding: quoted-printable We are getting ready to spec out a new machine and are wondering about the wisdom of buying a quad versus a dual processor machine. Seing as how postgres in not a threaded application, and this server will only be used for log/transaction analysis (it will only ever have a few large queries running). Is there any performance to be gained, and if so is it worth the large cost? Any thoughts/experience are much appreciated... --=20 Chris Field cfield@affinitysolutions.com Affinity Solutions Inc. 386 Park Avenue South Suite 1209 New York, NY 10016 (212) 685-8748 ext. 32 --=-Pp9U7L6ZAviF2yFBgYtQ Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.7 (GNU/Linux) iD8DBQA/sXGfkTUAUyBR2GYRAthQAJ9MUbHD426ykBbMWtbKc2udiPKXbwCfVDFY LHQzFgVUwTNDE8PQxHwSgfQ= =Z6sz -----END PGP SIGNATURE----- --=-Pp9U7L6ZAviF2yFBgYtQ-- From pgsql-performance-owner@postgresql.org Tue Nov 11 19:51:57 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EA3E7D1B4FA for ; Tue, 11 Nov 2003 23:51:55 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 56717-03 for ; Tue, 11 Nov 2003 19:51:27 -0400 (AST) Received: from mail.hive.nj2.inquent.com (mc.carriermail.com [205.178.180.9]) by svr1.postgresql.org (Postfix) with SMTP id 8ACD0D1B531 for ; Tue, 11 Nov 2003 19:51:23 -0400 (AST) Received: (qmail 4509 invoked from network); 11 Nov 2003 23:51:50 -0000 Received: from unknown (HELO ?192.168.1.199?) (134.22.68.129) by 205.178.180.9 with SMTP; 11 Nov 2003 23:51:50 -0000 Subject: Re: Value of Quad vs. Dual Processor machine From: Rod Taylor To: cfield@affinitysolutions.com Cc: Postgresql Performance In-Reply-To: <1068593567.2063.8.camel@tuxsrva> References: <1068593567.2063.8.camel@tuxsrva> Content-Type: text/plain Message-Id: <1068594688.28107.16.camel@jester> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Tue, 11 Nov 2003 18:51:28 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/52 X-Sequence-Number: 4612 On Tue, 2003-11-11 at 18:32, Chris Field wrote: > We are getting ready to spec out a new machine and are wondering about > the wisdom of buying a quad versus a dual processor machine. Seing as > how postgres in not a threaded application, and this server will only be > used for log/transaction analysis (it will only ever have a few large > queries running). Is there any performance to be gained, and if so is > it worth the large cost? Any thoughts/experience are much > appreciated... Since you're asking the question, I'll assume you don't have CPU intensive queries or monstrous loads. I'd probably invest in a Quad system with 2 chips in it (2 empty sockets) and put the difference in funds into a few extra GB of Ram or improved IO. In 6 months or a year, if you start doing longer or more complex queries, toss in the other 2 chips. So long as you don't hit a memory limit, it'll be fine. From pgsql-performance-owner@postgresql.org Tue Nov 11 20:55:10 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CD336D1B4FA for ; Wed, 12 Nov 2003 00:55:06 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 65173-10 for ; Tue, 11 Nov 2003 20:54:38 -0400 (AST) Received: from mail1.ihs.com (mail1.ihs.com [170.207.70.222]) by svr1.postgresql.org (Postfix) with ESMTP id BBD6CD1B50C for ; Tue, 11 Nov 2003 20:54:34 -0400 (AST) Received: from css120.ihs.com (css120.ihs.com [170.207.105.120]) by mail1.ihs.com (8.12.10/8.12.10) with ESMTP id hAC0sJNK002597; Tue, 11 Nov 2003 17:54:20 -0700 (MST) Date: Tue, 11 Nov 2003 17:40:14 -0700 (MST) From: "scott.marlowe" To: Rod Taylor Cc: , Postgresql Performance Subject: Re: Value of Quad vs. Dual Processor machine In-Reply-To: <1068594688.28107.16.camel@jester> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-IHS-MailScanner: Found to be clean X-IHS-MailScanner-SpamCheck: X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/53 X-Sequence-Number: 4613 On Tue, 11 Nov 2003, Rod Taylor wrote: > On Tue, 2003-11-11 at 18:32, Chris Field wrote: > > We are getting ready to spec out a new machine and are wondering about > > the wisdom of buying a quad versus a dual processor machine. Seing as > > how postgres in not a threaded application, and this server will only be > > used for log/transaction analysis (it will only ever have a few large > > queries running). Is there any performance to be gained, and if so is > > it worth the large cost? Any thoughts/experience are much > > appreciated... > > Since you're asking the question, I'll assume you don't have CPU > intensive queries or monstrous loads. > > I'd probably invest in a Quad system with 2 chips in it (2 empty > sockets) and put the difference in funds into a few extra GB of Ram or > improved IO. > > In 6 months or a year, if you start doing longer or more complex > queries, toss in the other 2 chips. So long as you don't hit a memory > limit, it'll be fine. Note that you want to carefully look at the difference in cost of the motherboard versus the CPUs. It's often the motherboard that raises the cost, not the CPUs so much. Although with Xeons, the CPUs are not cheap. The second issue is that Intel (and AMD probably) only guarantee proper performance from chips int he same batch, so you may wind up replacing the two working CPUs with two new ones to go with the other two you'll be buying, to make sure that they work together. My guess is that more CPUs aren't gonna help this problem a lot, so look more at fast RAM and lots of it, as well as a fast I/O subsystem. 2 CPUs should be plenty. From pgsql-performance-owner@postgresql.org Tue Nov 11 21:03:19 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0B941D1B51E for ; Wed, 12 Nov 2003 01:03:18 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 65104-10 for ; Tue, 11 Nov 2003 21:02:49 -0400 (AST) Received: from lifeintegrity.com (h000476f4f1ab.ne.client2.attbi.com [66.30.209.218]) by svr1.postgresql.org (Postfix) with ESMTP id 8CD7ED1B4FA for ; Tue, 11 Nov 2003 21:02:39 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by lifeintegrity.com (Postfix) with ESMTP id CAC2C60041 for ; Tue, 11 Nov 2003 20:02:41 -0500 (EST) Received: from lifeintegrity.com ([127.0.0.1]) by localhost (pawan.lifeintegrity.com [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 05712-02-10 for ; Tue, 11 Nov 2003 20:02:41 -0500 (EST) Received: by lifeintegrity.com (Postfix, from userid 1000) id 08A756001B; Tue, 11 Nov 2003 20:02:40 -0500 (EST) Date: Tue, 11 Nov 2003 20:02:40 -0500 To: Postgresql Performance Subject: Re: Value of Quad vs. Dual Processor machine Message-ID: <20031112010240.GA3751@lifeintegrity.com> Mail-Followup-To: Postgresql Performance References: <1068594688.28107.16.camel@jester> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="+QahgC5+KEYLbs62" Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.4i From: allanwind@lifeintegrity.com (Allan Wind) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/54 X-Sequence-Number: 4614 --+QahgC5+KEYLbs62 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On 2003-11-11T17:40:14-0700, scott.marlowe wrote: > 2 CPUs should be plenty. for everyone? No, I must have been thinking of someone else :-) /Allan --=20 Allan Wind P.O. Box 2022 Woburn, MA 01888-0022 USA --+QahgC5+KEYLbs62 Content-Type: application/pgp-signature; name="signature.asc" Content-Description: Digital signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.3 (GNU/Linux) iD8DBQE/sYawuDtNyOwreTYRAoe2AKDAQYKoQM7/5ZkDge18dnW6Gm2jRgCePedQ bKNjnJvKqwAhc1fK3bsGJ5A= =qskP -----END PGP SIGNATURE----- --+QahgC5+KEYLbs62-- From pgsql-performance-owner@postgresql.org Tue Nov 11 21:11:54 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9B9CED1B4F6 for ; Wed, 12 Nov 2003 01:11:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 65807-10 for ; Tue, 11 Nov 2003 21:11:23 -0400 (AST) Received: from redhotpenguin.com (12-240-9-202.client.attbi.com [12.240.9.202]) by svr1.postgresql.org (Postfix) with ESMTP id 31835D1B521 for ; Tue, 11 Nov 2003 21:11:19 -0400 (AST) Received: (qmail 29762 invoked from network); 12 Nov 2003 01:11:20 -0000 Received: from localhost (HELO 127.0.0.1) (127.0.0.1) by localhost with SMTP; 12 Nov 2003 01:11:20 -0000 Received: from 127.0.0.1 (proxying for 12.240.9.202) (SquirrelMail authenticated user fred@redhotpenguin.com) by 127.0.0.1 with HTTP; Tue, 11 Nov 2003 17:11:20 -0800 (PST) Message-ID: <53970.127.0.0.1.1068599480.squirrel@127.0.0.1> In-Reply-To: References: <1068594688.28107.16.camel@jester> Date: Tue, 11 Nov 2003 17:11:20 -0800 (PST) Subject: Re: Value of Quad vs. Dual Processor machine From: fred@redhotpenguin.com To: "scott.marlowe" Cc: "Rod Taylor" , cfield@affinitysolutions.com, "Postgresql Performance" User-Agent: SquirrelMail/1.4.2-1.qvcs.1 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/55 X-Sequence-Number: 4615 > On Tue, 11 Nov 2003, Rod Taylor wrote: > >> On Tue, 2003-11-11 at 18:32, Chris Field wrote: >> > We are getting ready to spec out a new machine and are wondering about >> > the wisdom of buying a quad versus a dual processor machine. Seing as >> > how postgres in not a threaded application, and this server will only >> be >> > used for log/transaction analysis (it will only ever have a few large >> > queries running). Is there any performance to be gained, and if so is >> > it worth the large cost? Any thoughts/experience are much >> > appreciated... >> >> Since you're asking the question, I'll assume you don't have CPU >> intensive queries or monstrous loads. >> >> I'd probably invest in a Quad system with 2 chips in it (2 empty >> sockets) and put the difference in funds into a few extra GB of Ram or >> improved IO. >> >> In 6 months or a year, if you start doing longer or more complex >> queries, toss in the other 2 chips. So long as you don't hit a memory >> limit, it'll be fine. > > Note that you want to carefully look at the difference in cost of the > motherboard versus the CPUs. It's often the motherboard that raises the > cost, not the CPUs so much. Although with Xeons, the CPUs are not cheap. > > The second issue is that Intel (and AMD probably) only guarantee proper > performance from chips int he same batch, so you may wind up replacing the > two working CPUs with two new ones to go with the other two you'll be > buying, to make sure that they work together. > > My guess is that more CPUs aren't gonna help this problem a lot, so look > more at fast RAM and lots of it, as well as a fast I/O subsystem. > > 2 CPUs should be plenty. I agree that the additional cpus won't help as much since I haven't found any benefits in terms of individual query speed for a quad vs. an smp on benchmarks I've run on test machines I was considering purchasing. Quads are also expensive - on similar architectures the quad was 20k vs 7k for the dual. > > > ---------------------------(end of broadcast)--------------------------- > TIP 2: you can get off all lists at once with the unregister command > (send "unregister YourEmailAddressHere" to majordomo@postgresql.org) > From pgsql-performance-owner@postgresql.org Tue Nov 11 21:25:23 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E5EDFD1B4F9 for ; Wed, 12 Nov 2003 01:25:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 74568-04 for ; Tue, 11 Nov 2003 21:24:52 -0400 (AST) Received: from lakemtao01.cox.net (lakemtao01.cox.net [68.1.17.244]) by svr1.postgresql.org (Postfix) with ESMTP id A87BDD1B4F6 for ; Tue, 11 Nov 2003 21:24:48 -0400 (AST) Received: from lhosts ([68.11.66.83]) by lakemtao01.cox.net (InterMail vM.5.01.06.05 201-253-122-130-105-20030824) with ESMTP id <20031112012451.LKBW23168.lakemtao01.cox.net@lhosts> for ; Tue, 11 Nov 2003 20:24:51 -0500 Subject: Re: Value of Quad vs. Dual Processor machine From: Ron Johnson To: PgSQL Performance ML In-Reply-To: <1068593567.2063.8.camel@tuxsrva> References: <1068593567.2063.8.camel@tuxsrva> Content-Type: text/plain Message-Id: <1068600291.30969.53.camel@haggis.homelan> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Tue, 11 Nov 2003 19:24:51 -0600 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/56 X-Sequence-Number: 4616 On Tue, 2003-11-11 at 17:32, Chris Field wrote: > We are getting ready to spec out a new machine and are wondering about > the wisdom of buying a quad versus a dual processor machine. Seing as > how postgres in not a threaded application, and this server will only be > used for log/transaction analysis (it will only ever have a few large > queries running). Is there any performance to be gained, and if so is > it worth the large cost? Any thoughts/experience are much > appreciated... Xeon or Opteron? The faster Opterons *really* blaze, especially in 64-bit mode. As others have said, though, RAM and I/O are most important. -- ----------------------------------------------------------------- Ron Johnson, Jr. ron.l.johnson@cox.net Jefferson, LA USA "As I like to joke, I may have invented it, but Microsoft made it popular" David Bradley, regarding Ctrl-Alt-Del From pgsql-performance-owner@postgresql.org Tue Nov 11 22:12:11 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5DF94D1B50B for ; Wed, 12 Nov 2003 02:12:09 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 75635-09 for ; Tue, 11 Nov 2003 22:11:41 -0400 (AST) Received: from corp1.affinitysolutions.com (corp1.affinitysolutions.com [216.52.88.186]) by svr1.postgresql.org (Postfix) with ESMTP id 76B2BD1B53B for ; Tue, 11 Nov 2003 22:11:37 -0400 (AST) Received: from win2000 (24-193-100-224.nyc.rr.com [24.193.100.224]) by corp1.affinitysolutions.com (Postfix) with ESMTP id BAAF111F703; Tue, 11 Nov 2003 21:11:39 -0500 (EST) Message-ID: <001401c3a8c2$8c8fec30$040210ac@win2000> Reply-To: "Chris Field" From: "Chris Field" To: "Ron Johnson" , "PgSQL Performance ML" References: <1068593567.2063.8.camel@tuxsrva> <1068600291.30969.53.camel@haggis.homelan> Subject: Re: Value of Quad vs. Dual Processor machine Date: Tue, 11 Nov 2003 21:13:19 -0500 Organization: Affinity Solutions MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.3 tagged_above=0.0 required=5.0 tests=BAYES_30, QUOTED_EMAIL_TEXT, RCVD_IN_NJABL, REFERENCES, X_NJABL_OPEN_PROXY X-Spam-Level: X-Archive-Number: 200311/57 X-Sequence-Number: 4617 we are looking at Xeon, We are currently running it on a quad sun v880 compiled to be 64bit and have been getting dreadful performance. I don't think we really have much to gain from going 64bit. ----- Original Message ----- From: "Ron Johnson" To: "PgSQL Performance ML" Sent: Tuesday, November 11, 2003 8:24 PM Subject: Re: [PERFORM] Value of Quad vs. Dual Processor machine > On Tue, 2003-11-11 at 17:32, Chris Field wrote: > > We are getting ready to spec out a new machine and are wondering about > > the wisdom of buying a quad versus a dual processor machine. Seing as > > how postgres in not a threaded application, and this server will only be > > used for log/transaction analysis (it will only ever have a few large > > queries running). Is there any performance to be gained, and if so is > > it worth the large cost? Any thoughts/experience are much > > appreciated... > > Xeon or Opteron? The faster Opterons *really* blaze, especially > in 64-bit mode. As others have said, though, RAM and I/O are most > important. > > -- > ----------------------------------------------------------------- > Ron Johnson, Jr. ron.l.johnson@cox.net > Jefferson, LA USA > > "As I like to joke, I may have invented it, but Microsoft made it > popular" > David Bradley, regarding Ctrl-Alt-Del > > > ---------------------------(end of broadcast)--------------------------- > TIP 5: Have you checked our extensive FAQ? > > http://www.postgresql.org/docs/faqs/FAQ.html > > From pgsql-performance-owner@postgresql.org Tue Nov 11 23:18:08 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 92EF5D1B536 for ; Wed, 12 Nov 2003 03:18:06 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 77815-10 for ; Tue, 11 Nov 2003 23:17:39 -0400 (AST) Received: from redhotpenguin.com (12-240-9-202.client.attbi.com [12.240.9.202]) by svr1.postgresql.org (Postfix) with ESMTP id D7DD7D1B4F9 for ; Tue, 11 Nov 2003 23:17:34 -0400 (AST) Received: (qmail 30551 invoked from network); 12 Nov 2003 03:17:38 -0000 Received: from 12-240-9-202.client.attbi.com (HELO harpua) (fred@redhotpenguin.com@12.240.9.202) by 192.168.0.2 with RC4-MD5 encrypted SMTP; 12 Nov 2003 03:17:38 -0000 From: "Fred Moyer" To: "'Chris Field'" , "'Ron Johnson'" , "'PgSQL Performance ML'" Subject: Re: Value of Quad vs. Dual Processor machine Date: Tue, 11 Nov 2003 19:17:40 -0800 Message-ID: <000001c3a8cb$87ff7880$0300a8c0@harpua> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.2616 In-Reply-To: <001401c3a8c2$8c8fec30$040210ac@win2000> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/58 X-Sequence-Number: 4618 One thing I learned after spending about a week comparing the Athlon (2 ghz, 333 mhz frontside bus) and Xeon (2.4 ghz, 266 mhz frontside bus) platforms was that on average the select queries I was benchmarking ran 30% faster on the Athlon (this was with data cached in memory so may not apply to the larger data sets where I/O is the limiting factor.) I benchmarked against the Opteron 244 when it came out and it came in about the same as the Athlon (makes sense since both were 333 mhz memory). The results within +/- 5-10% that of the Athlon. From testing against a couple of other machines I noticed that the memory bus speeds were almost directly proportional to the query times under these conditions. Not sure how these compare against the quad sun but the AMD chips returned the select queries faster than the Xeons from the informal investigations I did. Definitely try it before you buy it if possible. -----Original Message----- From: pgsql-performance-owner@postgresql.org [mailto:pgsql-performance-owner@postgresql.org] On Behalf Of Chris Field Sent: Tuesday, November 11, 2003 6:13 PM To: Ron Johnson; PgSQL Performance ML Subject: Re: [PERFORM] Value of Quad vs. Dual Processor machine we are looking at Xeon, We are currently running it on a quad sun v880 compiled to be 64bit and have been getting dreadful performance. I don't think we really have much to gain from going 64bit. ----- Original Message ----- From: "Ron Johnson" To: "PgSQL Performance ML" Sent: Tuesday, November 11, 2003 8:24 PM Subject: Re: [PERFORM] Value of Quad vs. Dual Processor machine > On Tue, 2003-11-11 at 17:32, Chris Field wrote: > > We are getting ready to spec out a new machine and are wondering > > about the wisdom of buying a quad versus a dual processor machine. > > Seing as how postgres in not a threaded application, and this server > > will only be used for log/transaction analysis (it will only ever > > have a few large queries running). Is there any performance to be > > gained, and if so is it worth the large cost? Any > > thoughts/experience are much appreciated... > > Xeon or Opteron? The faster Opterons *really* blaze, especially in > 64-bit mode. As others have said, though, RAM and I/O are most > important. > > -- > ----------------------------------------------------------------- > Ron Johnson, Jr. ron.l.johnson@cox.net > Jefferson, LA USA > > "As I like to joke, I may have invented it, but Microsoft made it > popular" David Bradley, regarding Ctrl-Alt-Del > > > ---------------------------(end of > broadcast)--------------------------- > TIP 5: Have you checked our extensive FAQ? > > http://www.postgresql.org/docs/faqs/FAQ.html > > ---------------------------(end of broadcast)--------------------------- TIP 7: don't forget to increase your free space map settings From pgsql-performance-owner@postgresql.org Wed Nov 12 01:53:18 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 47E97D1B508; Wed, 12 Nov 2003 05:53:17 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 18988-09; Wed, 12 Nov 2003 01:52:46 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 1D454D1B50C; Wed, 12 Nov 2003 01:52:45 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id 6550E3692A; Wed, 12 Nov 2003 00:52:44 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1AJnvo-0006jW-00; Wed, 12 Nov 2003 00:52:44 -0500 To: "Marc G. Fournier" Cc: Dennis Bjorklund , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... References: <20031111134925.L56037@ganymede.hub.org> In-Reply-To: <20031111134925.L56037@ganymede.hub.org> From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 12 Nov 2003 00:52:44 -0500 Message-ID: <87oevii15f.fsf@stark.dyndns.tv> Lines: 49 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/59 X-Sequence-Number: 4619 "Marc G. Fournier" writes: > Just as a side note, just doing a straight scan for the records, with no > SUM()/GROUP BY involved, with the month_trunc() index is still >8k msec: Well so the problem isn't the query at all, you just have too much data to massage online. You can preprocess the data offline into a more managable amount of data for your online reports. What I used to do for a similar situation was to do hourly queries sort of like this: insert into data_aggregate (day, hour, company_id, total_bytes) (select trunc(now(),'day'), trunc(now(), 'hour'), company_id, sum(bytes) from raw_data where time between trunc(now(),'hour') and trunc(now(),'hour')+'1 hour'::interval group by company_id ) [this was actually on oracle and the data looked kind of different, i'm making this up as i go along] Then later the reports could run quickly based on data_aggregate instead of slowly based on the much larger data set accumulated by the minute. Once I had this schema set up it was easy to follow it for all of the rapidly growing data tables. Now in my situation I had thousands of records accumulating per second, so hourly was already a big win. I originally chose hourly because I thought I might want time-of-day reports but that never panned out. On the other hand it was a win when the system broke once because I could easily see that and fix it before midnight when it would have actually mattered. Perhaps in your situation you would want daily aggregates or something else. One of the other advantages of these aggregate tables was that we could purge the old data much sooner with much less resistance from the business. Since the reports were all still available and a lot of ad-hoc queries could still be done without the raw data anyways. Alternatively you can just give up on online reports. Eventually you'll have some query that takes way more than 8s anyways. You can pregenerate the entire report as a batch job instead. Either send it off as a nightly e-mail, store it as an html or csv file for the web server, or (my favourite) store the data for the report as an sql table and then have multiple front-ends that do a simple "select *" to pull the data and format it. -- greg From pgsql-performance-owner@postgresql.org Wed Nov 12 04:43:42 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E0EBDD1B509 for ; Wed, 12 Nov 2003 08:43:38 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 36213-09 for ; Wed, 12 Nov 2003 04:43:08 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id B1F25D1B531 for ; Wed, 12 Nov 2003 04:43:06 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Wed, 12 Nov 2003 01:43:06 -0700 Message-ID: <3FB1F28E.702@myrealbox.com> Date: Wed, 12 Nov 2003 14:12:54 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Fred Moyer Cc: 'Chris Field' , 'Ron Johnson' , 'PgSQL Performance ML' Subject: Re: Value of Quad vs. Dual Processor machine References: <000001c3a8cb$87ff7880$0300a8c0@harpua> In-Reply-To: <000001c3a8cb$87ff7880$0300a8c0@harpua> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/60 X-Sequence-Number: 4620 Fred Moyer wrote: > One thing I learned after spending about a week comparing the Athlon (2 > ghz, 333 mhz frontside bus) and Xeon (2.4 ghz, 266 mhz frontside bus) > platforms was that on average the select queries I was benchmarking ran > 30% faster on the Athlon (this was with data cached in memory so may not > apply to the larger data sets where I/O is the limiting factor.) > > I benchmarked against the Opteron 244 when it came out and it came in > about the same as the Athlon (makes sense since both were 333 mhz > memory). The results within +/- 5-10% that of the Athlon. From testing > against a couple of other machines I noticed that the memory bus speeds > were almost directly proportional to the query times under these > conditions. I remember a posting here about opteron, which essentially said, even if opteron works on par with athlon under few clients, as load increases it scales more than 50% better than athlons. So that could be another shot at it.Sorry, no handy URL here. Shridhar From pgsql-performance-owner@postgresql.org Wed Nov 12 09:35:42 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D9B8FD1B51E for ; Wed, 12 Nov 2003 13:35:36 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 76401-08 for ; Wed, 12 Nov 2003 09:35:09 -0400 (AST) Received: from shire.ontko.com (shire.ontko.com [199.164.165.1]) by svr1.postgresql.org (Postfix) with ESMTP id 8FB63D1B51A for ; Wed, 12 Nov 2003 09:35:05 -0400 (AST) Received: from nick (bilbo.ontko.com [199.164.165.101]) by shire.ontko.com (8.12.3/8.12.3/Debian-6.6) with SMTP id hACDZ15f012637 for ; Wed, 12 Nov 2003 08:35:04 -0500 Reply-To: From: "Nick Fankhauser" To: "Pgsql-Performance@Postgresql. Org" Subject: Seeking help with a query that takes too long Date: Wed, 12 Nov 2003 08:34:50 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/61 X-Sequence-Number: 4621 [I originally posted this using the wrong E-Mail account, so a double posting may occur if the first message gets released by the moderator later- sorry!] Hi- I have a query that I'm trying to speed up. I haven't been able to come up with any workable ideas for speeding it up, so I'm seeking some input from the list. I'm using version 7.3.2 I have three tables: case_data (1,947,386 rows) actor (3,385,669 rows) actor_case_assignment (8,668,650 rows) As the names imply, actor_case_assignment contains records that assign an actor to a case. Actors such as attorneys or judges may have many cases, while the average actor (we hope) only has one. What I'm trying to do is link these tables to get back a single row per actor that shows the actor's name, the number of cases that actor is assigned to, and if they only have one case, I want the public_id for that case. This means I have to do a group by to get the case count, but I'm then forced to use an aggregate function like max on the other fields. All of the fields ending in "_id" have unique indexes, and actor_full_name_uppercase is indexed. An analyze is done every night & the database is fairly stable in it's composition. Here's the select: select actor.actor_id, max(actor.actor_full_name), max(case_data.case_public_id), max(case_data.case_id), count(case_data.case_id) as case_count from actor, actor_case_assignment, case_data where actor.actor_full_name_uppercase like upper('sanders%') and actor.actor_id = actor_case_assignment.actor_id and case_data.case_id = actor_case_assignment.case_id group by actor.actor_id order by max(actor.actor_full_name), case_count desc limit 1000; Here's the explain analyze: QUERY PLAN ---------------------------------------------------------------------------- ---------------------------------------------------------------------------- -------------------------------------------------- Limit (cost=2214.71..2214.72 rows=1 width=115) (actual time=120034.61..120035.67 rows=1000 loops=1) -> Sort (cost=2214.71..2214.72 rows=1 width=115) (actual time=120034.60..120034.98 rows=1001 loops=1) Sort Key: max((actor.actor_full_name)::text), count(case_data.case_id) -> Aggregate (cost=2214.67..2214.70 rows=1 width=115) (actual time=119962.80..120011.49 rows=3456 loops=1) -> Group (cost=2214.67..2214.68 rows=2 width=115) (actual time=119962.76..119987.04 rows=5879 loops=1) -> Sort (cost=2214.67..2214.68 rows=2 width=115) (actual time=119962.74..119965.09 rows=5879 loops=1) Sort Key: actor.actor_id -> Nested Loop (cost=0.00..2214.66 rows=2 width=115) (actual time=59.05..119929.71 rows=5879 loops=1) -> Nested Loop (cost=0.00..2205.26 rows=3 width=76) (actual time=51.46..66089.04 rows=5882 loops=1) -> Index Scan using actor_full_name_uppercase on actor (cost=0.00..6.01 rows=1 width=42) (actual time=37.62..677.44 rows=3501 loops=1) Index Cond: ((actor_full_name_uppercase >= 'SANDERS'::character varying) AND (actor_full_name_uppercase < 'SANDERT'::character varying)) Filter: (actor_full_name_uppercase ~~ 'SANDERS%'::text) -> Index Scan using actor_case_assignment_actor_id on actor_case_assignment (cost=0.00..2165.93 rows=2666 width=34) (actual time=16.37..18.67 rows=2 loops=3501) Index Cond: ("outer".actor_id = actor_case_assignment.actor_id) -> Index Scan using case_data_case_id on case_data (cost=0.00..3.66 rows=1 width=39) (actual time=9.14..9.15 rows=1 loops=5882) Index Cond: (case_data.case_id = "outer".case_id) Total runtime: 120038.60 msec (17 rows) Any ideas? Thanks! -Nick --------------------------------------------------------------------- Nick Fankhauser nickf@doxpop.com Phone 1.765.965.7363 Fax 1.765.962.9788 doxpop - Court records at your fingertips - http://www.doxpop.com/ From pgsql-performance-owner@postgresql.org Wed Nov 12 10:28:52 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5CE44D1B51C for ; Wed, 12 Nov 2003 14:28:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 91126-09 for ; Wed, 12 Nov 2003 10:28:20 -0400 (AST) Received: from jefftrout.com (h00a0cc4084e5.ne.client2.attbi.com [24.128.241.68]) by svr1.postgresql.org (Postfix) with SMTP id 73FB4D1B50D for ; Wed, 12 Nov 2003 10:28:15 -0400 (AST) Received: (qmail 77237 invoked from network); 12 Nov 2003 14:28:21 -0000 Received: from localhost (HELO squeegit) (threshar@127.0.0.1) by localhost with SMTP; 12 Nov 2003 14:28:21 -0000 Date: Wed, 12 Nov 2003 09:28:07 -0500 From: Jeff To: "Chris Field" Cc: ron.l.johnson@cox.net, pgsql-performance@postgresql.org Subject: Re: Value of Quad vs. Dual Processor machine Message-Id: <20031112092807.14b18bf4.threshar@torgo.978.org> In-Reply-To: <001401c3a8c2$8c8fec30$040210ac@win2000> References: <1068593567.2063.8.camel@tuxsrva> <1068600291.30969.53.camel@haggis.homelan> <001401c3a8c2$8c8fec30$040210ac@win2000> X-Mailer: Sylpheed version 0.9.7 (GTK+ 1.2.10; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/62 X-Sequence-Number: 4622 On Tue, 11 Nov 2003 21:13:19 -0500 "Chris Field" wrote: > we are looking at Xeon, We are currently running it on a quad sun v880 > compiled to be 64bit and have been getting dreadful performance. I > don't think we really have much to gain from going 64bit. > > By chance, are you running 7.3.4 on that sun? If so, try this: export CFLAGS=-02 ./configure and rebuild PG. Before 7.4 PG was build with _no_ optimization on Solaris. Recompiling gives __HUGE__ (notice the underscores) performance gains. And onto the dual vs quad. PG will only use 1 cpu / connection / query. So if your machine iwll have 1-2 queries running at a time those other 2 proc's will sit around idling. However if you are going to have a bunch going, 4 cpus will be most useful. One of hte nicest things to do for PG is more ram and fast IO. It really loves those things. good luck -- Jeff Trout http://www.jefftrout.com/ http://www.stuarthamm.net/ From pgsql-performance-owner@postgresql.org Tue Nov 11 10:47:41 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9383DD1B53F for ; Tue, 11 Nov 2003 14:47:38 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 59023-02 for ; Tue, 11 Nov 2003 10:47:11 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id EA891D1B531 for ; Tue, 11 Nov 2003 10:47:06 -0400 (AST) Received: (qmail 2566 invoked from network); 11 Nov 2003 14:48:21 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 11 Nov 2003 14:48:21 -0000 Message-ID: <3FB247FE.4080808@trade-india.com> Date: Wed, 12 Nov 2003 20:17:26 +0530 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Suggestions for benchmarking 7.4RC2 against 7.3 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/38 X-Sequence-Number: 4598 Hi, I plan to put 7.4-RC2 in our production servers in next few hours. Since the hardware config & the performance related GUCs parameter are going to remain the same i am interested in seeing the performance improvements in 7.4 as compared 7.3 . For this i plan to use the OSDB 0.14 and compare the results for both the cases. Does any one has suggestions for comparing 7.4 against 7.3 ? Since i am using OSDB for second time only any tips/guidance on usage of that is also appreciated. H/W config: CPU: 4 X Intel(R) Xeon(TM) CPU 2.00GHz MEM : 2 GB I/O config : PGDATA on 10000 RPM Ultra160 scsi , pg_xlog on a similar seperate SCSI GUC: shared_buffers = 10000 max_fsm_relations = 5000 max_fsm_pages = 55099264 sort_mem = 16384 vacuum_mem = 8192 All other performance related parameter have default value eg: #effective_cache_size = 1000 # typically 8KB each #random_page_cost = 4 # units are one sequential page fetch cost #cpu_tuple_cost = 0.01 # (same) #cpu_index_tuple_cost = 0.001 # (same) #cpu_operator_cost = 0.0025 # (same) BTW i get following error at the moment: ----------------------------------------- /usr/local/bin/osdb-pg-ui --postgresql=no_hash_index "osdb" "Invoked: /usr/local/bin/osdb-pg-ui --postgresql=no_hash_index" create_tables() 0.78 seconds return value = 0 load() 1.02 seconds return value = 0 create_idx_uniques_key_bt() 0.64 seconds return value = 0 create_idx_updates_key_bt() 0.61 seconds return value = 0 create_idx_hundred_key_bt() 0.61 seconds return value = 0 create_idx_tenpct_key_bt() 0.62 seconds return value = 0 create_idx_tenpct_key_code_bt() 0.45 seconds return value = 0 create_idx_tiny_key_bt() 0.46 seconds return value = 0 create_idx_tenpct_int_bt() 0.46 seconds return value = 0 create_idx_tenpct_signed_bt() 0.45 seconds return value = 0 create_idx_uniques_code_h() 0.46 seconds return value = 0 create_idx_tenpct_double_bt() 0.46 seconds return value = 0 create_idx_updates_decim_bt() 0.45 seconds return value = 0 create_idx_tenpct_float_bt() 0.46 seconds return value = 0 create_idx_updates_int_bt() 0.46 seconds return value = 0 create_idx_tenpct_decim_bt() 0.46 seconds return value = 0 create_idx_hundred_code_h() 0.45 seconds return value = 0 create_idx_tenpct_name_h() 0.46 seconds return value = 0 create_idx_updates_code_h() 0.46 seconds return value = 0 create_idx_tenpct_code_h() 0.45 seconds return value = 0 create_idx_updates_double_bt() 0.46 seconds return value = 0 create_idx_hundred_foreign() 0.41 seconds return value = 0 populateDataBase() 11.54 seconds return value = 0 Error in test Counting tuples at (6746)osdb.c:294: ... empty database -- empty results perror() reports: Resource temporarily unavailable someone sighup'd the parent Any clue? ------------------------------------------ Regards Mallah. From pgsql-performance-owner@postgresql.org Wed Nov 12 11:06:32 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 96BBED1B4F2 for ; Wed, 12 Nov 2003 15:06:30 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 97969-07 for ; Wed, 12 Nov 2003 11:06:03 -0400 (AST) Received: from email06.aon.at (WARSL402PIP3.highway.telekom.at [195.3.96.75]) by svr1.postgresql.org (Postfix) with SMTP id 79A76D1B51C for ; Wed, 12 Nov 2003 11:05:58 -0400 (AST) Received: (qmail 104646 invoked from network); 12 Nov 2003 15:05:56 -0000 Received: from m161p002.dipool.highway.telekom.at (HELO cantor) ([62.46.10.2]) (envelope-sender ) by qmail6rs.highway.telekom.at (qmail-ldap-1.03) with SMTP for ; 12 Nov 2003 15:05:56 -0000 From: Manfred Koizar To: Cc: "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long Date: Wed, 12 Nov 2003 16:04:48 +0100 Message-ID: <03i4rv8cl6bgvr8r2c77ert3nfs0glfib8@email.aon.at> References: In-Reply-To: X-Mailer: Forte Agent 1.93/32.576 English (American) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/63 X-Sequence-Number: 4623 On Wed, 12 Nov 2003 08:34:50 -0500, "Nick Fankhauser" wrote: > -> Index Scan using >actor_full_name_uppercase on actor (cost=0.00..6.01 rows=1 width=42) ^^^^^^ >(actual time=37.62..677.44 rows=3501 loops=1) ^^^^^^^^^ > Index Cond: >((actor_full_name_uppercase >= 'SANDERS'::character varying) AND >(actor_full_name_uppercase < 'SANDERT'::character varying)) > Filter: >(actor_full_name_uppercase ~~ 'SANDERS%'::text) Nick, can you find out why this row count estimation is so far off? \x SELECT * FROM pg_stats WHERE tablename='actor' AND attname='actor_full_name_uppercase'; BTW, there seem to be missing cases: > -> Nested Loop (cost=0.00..2214.66 rows=2 width=115) > (actual time=59.05..119929.71 rows=5879 loops=1) ^^^^ > -> Nested Loop (cost=0.00..2205.26 rows=3 width=76) > (actual time=51.46..66089.04 rows=5882 loops=1) ^^^^ Servus Manfred From pgsql-performance-owner@postgresql.org Tue Nov 11 11:05:14 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4AB73D1B540 for ; Tue, 11 Nov 2003 15:05:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 59889-03 for ; Tue, 11 Nov 2003 11:04:45 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id D5230D1B535 for ; Tue, 11 Nov 2003 11:04:40 -0400 (AST) Received: (qmail 3778 invoked from network); 11 Nov 2003 15:05:55 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 11 Nov 2003 15:05:55 -0000 Message-ID: <3FB24C15.1020907@trade-india.com> Date: Wed, 12 Nov 2003 20:34:53 +0530 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Rajesh Kumar Mallah Cc: pgsql-performance@postgresql.org Subject: Re: Suggestions for benchmarking 7.4RC2 against 7.3 References: <3FB247FE.4080808@trade-india.com> In-Reply-To: <3FB247FE.4080808@trade-india.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/39 X-Sequence-Number: 4599 the error mentioned in first email has been overcome by running osdb on the same machine hosting the DB server. regds mallah. Rajesh Kumar Mallah wrote: > > Hi, > > I plan to put 7.4-RC2 in our production servers in next few hours. > > Since the hardware config & the performance related GUCs parameter > are going to remain the same i am interested in seeing the performance > improvements in 7.4 as compared 7.3 . > > For this i plan to use the OSDB 0.14 and compare the results for both > the > cases. > > Does any one has suggestions for comparing 7.4 against 7.3 ? > Since i am using OSDB for second time only any tips/guidance > on usage of that is also appreciated. > > > > H/W config: > > CPU: 4 X Intel(R) Xeon(TM) CPU 2.00GHz > MEM : 2 GB > I/O config : PGDATA on 10000 RPM Ultra160 scsi , pg_xlog on a similar > seperate SCSI > > GUC: > shared_buffers = 10000 > max_fsm_relations = 5000 > max_fsm_pages = 55099264 > sort_mem = 16384 > vacuum_mem = 8192 > > All other performance related parameter have default > value eg: > > #effective_cache_size = 1000 # typically 8KB each > #random_page_cost = 4 # units are one sequential page fetch > cost > #cpu_tuple_cost = 0.01 # (same) > #cpu_index_tuple_cost = 0.001 # (same) > #cpu_operator_cost = 0.0025 # (same) > > > > BTW i get following error at the moment: > ----------------------------------------- > /usr/local/bin/osdb-pg-ui --postgresql=no_hash_index > "osdb" > "Invoked: /usr/local/bin/osdb-pg-ui --postgresql=no_hash_index" > > create_tables() 0.78 seconds return value = 0 > load() 1.02 seconds return value = 0 > create_idx_uniques_key_bt() 0.64 seconds return value = 0 > create_idx_updates_key_bt() 0.61 seconds return value = 0 > create_idx_hundred_key_bt() 0.61 seconds return value = 0 > create_idx_tenpct_key_bt() 0.62 seconds return value = 0 > create_idx_tenpct_key_code_bt() 0.45 seconds return value = 0 > create_idx_tiny_key_bt() 0.46 seconds return value = 0 > create_idx_tenpct_int_bt() 0.46 seconds return value = 0 > create_idx_tenpct_signed_bt() 0.45 seconds return value = 0 > create_idx_uniques_code_h() 0.46 seconds return value = 0 > create_idx_tenpct_double_bt() 0.46 seconds return value = 0 > create_idx_updates_decim_bt() 0.45 seconds return value = 0 > create_idx_tenpct_float_bt() 0.46 seconds return value = 0 > create_idx_updates_int_bt() 0.46 seconds return value = 0 > create_idx_tenpct_decim_bt() 0.46 seconds return value = 0 > create_idx_hundred_code_h() 0.45 seconds return value = 0 > create_idx_tenpct_name_h() 0.46 seconds return value = 0 > create_idx_updates_code_h() 0.46 seconds return value = 0 > create_idx_tenpct_code_h() 0.45 seconds return value = 0 > create_idx_updates_double_bt() 0.46 seconds return value = 0 > create_idx_hundred_foreign() 0.41 seconds return value = 0 > populateDataBase() 11.54 seconds return value = 0 > > Error in test Counting tuples at (6746)osdb.c:294: > ... empty database -- empty results > perror() reports: Resource temporarily unavailable > > someone sighup'd the parent > > Any clue? > > ------------------------------------------ > > > Regards > Mallah. > > > > > > > ---------------------------(end of broadcast)--------------------------- > TIP 5: Have you checked our extensive FAQ? > > http://www.postgresql.org/docs/faqs/FAQ.html From pgsql-performance-owner@postgresql.org Wed Nov 12 11:44:17 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 27205D1B4FE for ; Wed, 12 Nov 2003 15:44:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 08763-08 for ; Wed, 12 Nov 2003 11:43:45 -0400 (AST) Received: from spirit.aldeiadigital.com.br (iplus-fac-225-213-137.xdsl-fixo.ctbcnetsuper.com.br [200.225.213.137]) by svr1.postgresql.org (Postfix) with ESMTP id 7574ED1B537 for ; Wed, 12 Nov 2003 11:43:35 -0400 (AST) Received: from webmail.ad2.com.br (localhost.localdomain [127.0.0.1]) by spirit.aldeiadigital.com.br (8.11.6/8.11.6) with SMTP id hACFf3715458 for ; Wed, 12 Nov 2003 13:41:03 -0200 Received: from 192.168.1.120 (SquirrelMail authenticated user alepaes) by webmail.ad2.com.br with HTTP; Wed, 12 Nov 2003 13:41:03 -0200 (BRST) Message-ID: <2834.192.168.1.120.1068651663.squirrel@webmail.ad2.com.br> Date: Wed, 12 Nov 2003 13:41:03 -0200 (BRST) Subject: Superior performance in PG 7.4 From: "alexandre :: aldeia digital" To: pgsql-performance@postgresql.org User-Agent: SquirrelMail/1.4.0 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 X-Priority: 3 Importance: Normal X-Virus-Scanned: by amavisd-milter (http://amavis.org/) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/64 X-Sequence-Number: 4624 Hi, I am trying the PG 7.4 RC1 and RC2 and I see a superb performance improvement compared with 7.3.... Explaining the querys, I see a change of planner that, in my case, prefer Nested Loops in 7.4 opposite to Hash or Merge Join in 7.3. To test, I disable Hash and Merge Joins in 7.3 and performance have been very improved using nested loops... Both systems are identical in configurations, properly vacuuned and analyzed before tests. Something can be wrong with my tests ? [ I desire that not :) ] Alexandre From pgsql-performance-owner@postgresql.org Wed Nov 12 12:10:05 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5C137D1B553 for ; Wed, 12 Nov 2003 16:10:01 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 11057-08 for ; Wed, 12 Nov 2003 12:09:30 -0400 (AST) Received: from mail1.ihs.com (mail1.ihs.com [170.207.70.222]) by svr1.postgresql.org (Postfix) with ESMTP id 86641D1B4EB for ; Wed, 12 Nov 2003 12:09:26 -0400 (AST) Received: from css120.ihs.com (css120.ihs.com [170.207.105.120]) by mail1.ihs.com (8.12.10/8.12.10) with ESMTP id hACG87NK009343; Wed, 12 Nov 2003 09:08:08 -0700 (MST) Date: Wed, 12 Nov 2003 08:53:56 -0700 (MST) From: "scott.marlowe" To: Chris Field Cc: Ron Johnson , PgSQL Performance ML Subject: Re: Value of Quad vs. Dual Processor machine In-Reply-To: <001401c3a8c2$8c8fec30$040210ac@win2000> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-IHS-MailScanner: Found to be clean X-IHS-MailScanner-SpamCheck: X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/67 X-Sequence-Number: 4627 As another post pointed out, you need to set cflags to get optimization under Solaris on that flavor of Postgresql. Also, Postgresql tends to get its best performance from the free unixes, Linux and BSD. those are available for Sun Sparcs, but postgresql in 64 bit mode on those boxes is still a bit cutting edge. It might be worth a try to set up the sun to dual boot to either BSD or Linux and test Postgresql under that environment to see how it works and compare it to Sun after you've set the cflags and recompiled. On Tue, 11 Nov 2003, Chris Field wrote: > we are looking at Xeon, We are currently running it on a quad sun v880 > compiled to be 64bit and have been getting dreadful performance. I don't > think we really have much to gain from going 64bit. > > > ----- Original Message ----- > From: "Ron Johnson" > To: "PgSQL Performance ML" > Sent: Tuesday, November 11, 2003 8:24 PM > Subject: Re: [PERFORM] Value of Quad vs. Dual Processor machine > > > > On Tue, 2003-11-11 at 17:32, Chris Field wrote: > > > We are getting ready to spec out a new machine and are wondering about > > > the wisdom of buying a quad versus a dual processor machine. Seing as > > > how postgres in not a threaded application, and this server will only be > > > used for log/transaction analysis (it will only ever have a few large > > > queries running). Is there any performance to be gained, and if so is > > > it worth the large cost? Any thoughts/experience are much > > > appreciated... > > > > Xeon or Opteron? The faster Opterons *really* blaze, especially > > in 64-bit mode. As others have said, though, RAM and I/O are most > > important. > > > > -- > > ----------------------------------------------------------------- > > Ron Johnson, Jr. ron.l.johnson@cox.net > > Jefferson, LA USA > > > > "As I like to joke, I may have invented it, but Microsoft made it > > popular" > > David Bradley, regarding Ctrl-Alt-Del > > > > > > ---------------------------(end of broadcast)--------------------------- > > TIP 5: Have you checked our extensive FAQ? > > > > http://www.postgresql.org/docs/faqs/FAQ.html > > > > > > > ---------------------------(end of broadcast)--------------------------- > TIP 7: don't forget to increase your free space map settings > > From pgsql-performance-owner@postgresql.org Wed Nov 12 11:56:34 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C1D67D1B4ED for ; Wed, 12 Nov 2003 15:56:30 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 11948-01 for ; Wed, 12 Nov 2003 11:56:02 -0400 (AST) Received: from shire.ontko.com (shire.ontko.com [199.164.165.1]) by svr1.postgresql.org (Postfix) with ESMTP id 4B67FD1B521 for ; Wed, 12 Nov 2003 11:55:58 -0400 (AST) Received: from nick (bilbo.ontko.com [199.164.165.101]) by shire.ontko.com (8.12.3/8.12.3/Debian-6.6) with SMTP id hACFtx5f015013; Wed, 12 Nov 2003 10:56:01 -0500 Reply-To: From: "Nick Fankhauser" To: "Manfred Koizar" , "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long Date: Wed, 12 Nov 2003 10:55:48 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) In-Reply-To: <03i4rv8cl6bgvr8r2c77ert3nfs0glfib8@email.aon.at> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.1 tagged_above=0.0 required=5.0 tests=IN_REP_TO, MSGID_GOOD_EXCHANGE, QUOTED_EMAIL_TEXT, UPPERCASE_25_50 X-Spam-Level: X-Archive-Number: 200311/65 X-Sequence-Number: 4625 >(actual time=37.62..677.44 rows=3501 loops=1) ^^^^^^^^^ > Nick, can you find out why this row count estimation is so far off? It's actually correct: prod1=# select count(actor_id) from actor where actor_full_name_uppercase like 'SANDERS%'; count ------- 3501 (1 row) Of course, I merely chose "SANDERS" arbitrarily as a name that falls somewhere near the middle of the frequency range for names. SMITH or JONES would represent a worst-case, and something like KOIZAR would probably be unique. Here are the stats: prod1=# SELECT * FROM pg_stats prod1-# WHERE tablename='actor' AND attname='actor_full_name_uppercase'; -[ RECORD 1 ]-----+------------------------------------------------------------------- ---------------------------------------------------------------------------- ---------------------------------------------------------------------------- ------------------------------------------------------------- schemaname | public tablename | actor attname | actor_full_name_uppercase null_frac | 0.000333333 avg_width | 21 n_distinct | 24215 most_common_vals | {"STATE OF INDIANA","INDIANA DEPARTMENT OF REVENUE","BARTH CONS SCHOOL CORP","HOWARD COUNTY CLERK","ADVANCED RECOVERY SERVICES","STATE OF INDIANA-DEPT OF REVENUE","ALLIED COLLECTION SERVICE INC","CREDIT BUREAU OF LAPORTE","MIDWEST COLLECTION SVC INC","NCO FINANCIAL SYSTEMS INC"} most_common_freqs | {0.0153333,0.0143333,0.00433333,0.00433333,0.004,0.00366667,0.00333333,0.003 33333,0.00266667,0.00266667} histogram_bounds | {"(POE) ESTELLE, DENISE","BRIEN, LIISI","COTTRELL, CAROL","FAMILY RENTALS","HAYNES, TAMIKA","KESSLER, VICTORIA","MEFFORD, VERNON L","PHILLIPS, GERALD L","SHELTON, ANTOINETTE","TRICARICO, MELISSA SUE","ZUEHLKE, THOMAS L"} correlation | -0.00147395 I think this means that the average is 357 per actor. As you can see, the range of assignments varies from people with a single parking ticket to "State of Indiana", which is party to many thousands of cases. > BTW, there seem to be missing cases: > > -> Nested Loop (cost=0.00..2214.66 rows=2 width=115) > > (actual time=59.05..119929.71 rows=5879 loops=1) > ^^^^ > > -> Nested Loop (cost=0.00..2205.26 rows=3 width=76) > > (actual time=51.46..66089.04 rows=5882 loops=1) This is expected- We actually aggregate data from many county court databases, with varying levels of data "cleanliness". Regards, -Nick From pgsql-performance-owner@postgresql.org Wed Nov 12 12:06:06 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E3BBAD1B534 for ; Wed, 12 Nov 2003 16:05:59 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 10607-09 for ; Wed, 12 Nov 2003 12:05:28 -0400 (AST) Received: from shire.ontko.com (shire.ontko.com [199.164.165.1]) by svr1.postgresql.org (Postfix) with ESMTP id 5875DD1B4EB for ; Wed, 12 Nov 2003 12:05:27 -0400 (AST) Received: from nick (bilbo.ontko.com [199.164.165.101]) by shire.ontko.com (8.12.3/8.12.3/Debian-6.6) with SMTP id hACG5M5f015295; Wed, 12 Nov 2003 11:05:23 -0500 Reply-To: From: "Nick Fankhauser" To: "Manfred Koizar" , "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long Date: Wed, 12 Nov 2003 11:05:10 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) In-Reply-To: <03i4rv8cl6bgvr8r2c77ert3nfs0glfib8@email.aon.at> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/66 X-Sequence-Number: 4626 > >actor_full_name_uppercase on actor (cost=0.00..6.01 rows=1 width=42) > ^^^^^^ > >(actual time=37.62..677.44 rows=3501 loops=1) > ^^^^^^^^^ > Nick, can you find out why this row count estimation is so far off? ^^^^^^^^^ Oops- I read this backward- I see what you mean now. That's a good question. I'm not sure what part of the stats this estimate might be pulled from. The average is 357, but the most common frequency may be around 1. -Nick From pgsql-performance-owner@postgresql.org Wed Nov 12 12:11:35 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 16359D1B560; Wed, 12 Nov 2003 16:11:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 11295-10; Wed, 12 Nov 2003 12:11:02 -0400 (AST) Received: from ganymede.hub.org (u46n208.hfx.eastlink.ca [24.222.46.208]) by svr1.postgresql.org (Postfix) with ESMTP id 424CAD1B550; Wed, 12 Nov 2003 12:11:02 -0400 (AST) Received: by ganymede.hub.org (Postfix, from userid 1000) id 282C5372DC; Wed, 12 Nov 2003 12:08:56 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by ganymede.hub.org (Postfix) with ESMTP id 274CC372DA; Wed, 12 Nov 2003 12:08:56 -0400 (AST) Date: Wed, 12 Nov 2003 12:08:56 -0400 (AST) From: "Marc G. Fournier" X-X-Sender: scrappy@ganymede.hub.org To: Greg Stark Cc: "Marc G. Fournier" , Dennis Bjorklund , pgsql-performance@postgresql.org Subject: Re: *very* slow query to summarize data for a month ... In-Reply-To: <87oevii15f.fsf@stark.dyndns.tv> Message-ID: <20031112120814.H56037@ganymede.hub.org> References: <20031111134925.L56037@ganymede.hub.org> <87oevii15f.fsf@stark.dyndns.tv> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/69 X-Sequence-Number: 4629 On Wed, 12 Nov 2003, Greg Stark wrote: > > "Marc G. Fournier" writes: > > > Just as a side note, just doing a straight scan for the records, with no > > SUM()/GROUP BY involved, with the month_trunc() index is still >8k msec: > > One of the other advantages of these aggregate tables was that we could > purge the old data much sooner with much less resistance from the > business. Since the reports were all still available and a lot of ad-hoc > queries could still be done without the raw data anyways. Actually, what I've done is do this at the 'load stage' ... but same concept ... From pgsql-performance-owner@postgresql.org Wed Nov 12 12:11:18 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4F514D1B55D for ; Wed, 12 Nov 2003 16:11:16 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 11269-09 for ; Wed, 12 Nov 2003 12:10:44 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 27CA6D1B50D for ; Wed, 12 Nov 2003 12:10:44 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hACGAY19009084; Wed, 12 Nov 2003 11:10:35 -0500 (EST) To: nickf@ontko.com Cc: "Manfred Koizar" , "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long In-reply-to: References: Comments: In-reply-to "Nick Fankhauser" message dated "Wed, 12 Nov 2003 10:55:48 -0500" Date: Wed, 12 Nov 2003 11:10:34 -0500 Message-ID: <9083.1068653434@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/68 X-Sequence-Number: 4628 "Nick Fankhauser" writes: >> Nick, can you find out why this row count estimation is so far off? > It's actually correct: Sure, the 3501 was the "actual". The estimate was 1 row, which was pretty far off :-( > Here are the stats: It looks like you are running with the default statistics target (10). Try boosting it to 100 or even more for this column (see ALTER TABLE SET STATISTICS, then re-ANALYZE) and see if the estimate gets better. I think the major problem is likely here: > n_distinct | 24215 which is no doubt much too small (do you have an idea of the number of distinct actor_full_name_uppercase values?) regards, tom lane From pgsql-performance-owner@postgresql.org Wed Nov 12 12:53:51 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 382B0D1B50C for ; Wed, 12 Nov 2003 16:53:49 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 22780-01 for ; Wed, 12 Nov 2003 12:53:18 -0400 (AST) Received: from shire.ontko.com (shire.ontko.com [199.164.165.1]) by svr1.postgresql.org (Postfix) with ESMTP id CFB6DD1B50B for ; Wed, 12 Nov 2003 12:53:17 -0400 (AST) Received: from nick (bilbo.ontko.com [199.164.165.101]) by shire.ontko.com (8.12.3/8.12.3/Debian-6.6) with SMTP id hACGr25f016224; Wed, 12 Nov 2003 11:53:04 -0500 Reply-To: From: "Nick Fankhauser" To: "Tom Lane" Cc: "Manfred Koizar" , "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long Date: Wed, 12 Nov 2003 11:52:51 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) In-Reply-To: <9083.1068653434@sss.pgh.pa.us> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/70 X-Sequence-Number: 4630 > It looks like you are running with the default statistics target (10). > Try boosting it to 100 or even more for this column (see ALTER TABLE > SET STATISTICS, then re-ANALYZE) and see if the estimate gets better. Here are the results & a few more clues: prod1=# alter table actor alter column actor_full_name_uppercase set statistics 1000; ALTER TABLE prod1=# analyze actor; ANALYZE prod1=# select count(distinct actor_full_name_uppercase) from actor; count --------- 1453371 (1 row) prod1=# select count(actor_id) from actor; count --------- 3386359 (1 row) This indicates to me that 1 isn't too shabby as an estimate if the whole name is specified, but I'm not sure how this gets altered in the case of a "LIKE" prod1=# \x Expanded display is on. prod1=# SELECT * FROM pg_stats prod1-# WHERE tablename='actor' AND attname='actor_full_name_uppercase';
schemaname | public tablename | actor attname | actor_full_name_uppercase null_frac | 0.000586667 avg_width | 21 n_distinct | -0.14701 correlation | -0.00211291 Question: What does it mean when n_distinct is negative? New results of explain analyze: QUERY PLAN ---------------------------------------------------------------------------- ---------------------------------------------------------------------------- -------------------------------------------------- Limit (cost=252683.61..252683.68 rows=28 width=116) (actual time=169377.32..169378.39 rows=1000 loops=1) -> Sort (cost=252683.61..252683.68 rows=29 width=116) (actual time=169377.31..169377.69 rows=1001 loops=1) Sort Key: max((actor.actor_full_name)::text), count(case_data.case_id) -> Aggregate (cost=252678.57..252682.91 rows=29 width=116) (actual time=169305.79..169354.50 rows=3456 loops=1) -> Group (cost=252678.57..252680.01 rows=289 width=116) (actual time=169305.76..169330.00 rows=5879 loops=1) -> Sort (cost=252678.57..252679.29 rows=289 width=116) (actual time=169305.75..169308.15 rows=5879 loops=1) Sort Key: actor.actor_id -> Nested Loop (cost=0.00..252666.74 rows=289 width=116) (actual time=89.27..169273.51 rows=5879 loops=1) -> Nested Loop (cost=0.00..251608.11 rows=289 width=77) (actual time=57.73..92753.49 rows=5882 loops=1) -> Index Scan using actor_full_name_uppercase on actor (cost=0.00..456.88 rows=113 width=42) (actual time=32.80..3197.28 rows=3501 loops=1) Index Cond: ((actor_full_name_uppercase >= 'SANDERS'::character varying) AND (actor_full_name_uppercase < 'SANDERT'::character varying)) Filter: (actor_full_name_uppercase ~~ 'SANDERS%'::text) -> Index Scan using actor_case_assignment_actor_id on actor_case_assignment (cost=0.00..2181.29 rows=2616 width=35) (actual time=22.26..25.57 rows=2 loops=3501) Index Cond: ("outer".actor_id = actor_case_assignment.actor_id) -> Index Scan using case_data_case_id on case_data (cost=0.00..3.65 rows=1 width=39) (actual time=13.00..13.00 rows=1 loops=5882) Index Cond: (case_data.case_id = "outer".case_id) Total runtime: 169381.38 msec (17 rows) From pgsql-performance-owner@postgresql.org Wed Nov 12 13:11:11 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5C793D1B502 for ; Wed, 12 Nov 2003 17:11:10 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 22209-06 for ; Wed, 12 Nov 2003 13:10:39 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 4F85CD1B4F2 for ; Wed, 12 Nov 2003 13:10:38 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hACHAa19010145; Wed, 12 Nov 2003 12:10:36 -0500 (EST) To: nickf@ontko.com Cc: "Manfred Koizar" , "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long In-reply-to: References: Comments: In-reply-to "Nick Fankhauser" message dated "Wed, 12 Nov 2003 11:52:51 -0500" Date: Wed, 12 Nov 2003 12:10:35 -0500 Message-ID: <10144.1068657035@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/71 X-Sequence-Number: 4631 "Nick Fankhauser" writes: > This indicates to me that 1 isn't too shabby as an estimate if the whole > name is specified, but I'm not sure how this gets altered in the case of a > "LIKE" For a pattern like "SANDERS%", the estimate is basically a range estimate for this condition: > ((actor_full_name_uppercase >= 'SANDERS'::character varying) AND > (actor_full_name_uppercase < 'SANDERT'::character varying)) > n_distinct | -0.14701 > Question: What does it mean when n_distinct is negative? It means that the number of distinct values is estimated as a fraction of the table size, rather than an absolute number. In this case 14.7% of the table size, which is a bit off compared to the correct value of 43% (1453371/3386359), but at least it's of the right order of magnitude now ... > -> Index Scan using > actor_full_name_uppercase on actor (cost=0.00..456.88 rows=113 width=42) > (actual time=32.80..3197.28 rows=3501 loops=1) Hmm. Better, but not enough better to force a different plan choice. You might have to resort to brute force, like "set enable_nestloop=false". Just out of curiosity, what do you get if you do that? regards, tom lane From pgsql-performance-owner@postgresql.org Wed Nov 12 13:38:43 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A6FF2D1B4FA for ; Wed, 12 Nov 2003 17:38:42 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 23157-08 for ; Wed, 12 Nov 2003 13:38:12 -0400 (AST) Received: from node-4024d1a2.mdw.onnet.us.uu.net (nsit-s229-45.uchicago.edu [128.135.229.45]) by svr1.postgresql.org (Postfix) with ESMTP id BF669D1B50C for ; Wed, 12 Nov 2003 13:38:10 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by node-4024d1a2.mdw.onnet.us.uu.net (Postfix) with ESMTP id 1E7D337D0A for ; Wed, 12 Nov 2003 11:34:42 -0600 (CST) Subject: performance optimzations From: Suchandra Thapa To: pgsql-performance@postgresql.org Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-P64FTYF3MjJ2djem0LYO" Organization: Message-Id: <1068658430.1071.42.camel@hepcat> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.2.2 (1.2.2-5) Date: 12 Nov 2003 11:34:41 -0600 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/72 X-Sequence-Number: 4632 --=-P64FTYF3MjJ2djem0LYO Content-Type: text/plain Content-Transfer-Encoding: quoted-printable I'm moving a webmail service over to use a postgresql database for storage and wanted to get any tips for optimizing performance. The machine will be a multiprocessor (either 2 or 4 cpu ) system with a raid array. What layout should be used? I was thinking using about using a raid 1+0 array to hold the database but since I can use different array types, would it be better to use 1+0 for the wal logs and a raid 5 for the database? The database gets fairly heavy activity (the system handles about 500MB of incoming and about 750MB of outgoing emails daily). I have a fairly free rein in regards to the system's layout as well as how the applications will interact with the database since I'm writing the code.=20=20 --=20 Suchandra Thapa --=-P64FTYF3MjJ2djem0LYO Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQA/sm7+6nShCjt5AZIRAh79AKCmQ3R4/p+myiLn8aRSxSDSTKdBzwCfaSgK gAxvE8cnf1azIdkoECAsO7s= =gNB3 -----END PGP SIGNATURE----- --=-P64FTYF3MjJ2djem0LYO-- From pgsql-performance-owner@postgresql.org Wed Nov 12 14:23:33 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 24D47D1B565 for ; Wed, 12 Nov 2003 18:23:31 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 39874-06 for ; Wed, 12 Nov 2003 14:23:00 -0400 (AST) Received: from mail.hive.nj2.inquent.com (mc.carriermail.com [205.178.180.9]) by svr1.postgresql.org (Postfix) with SMTP id D734BD1B561 for ; Wed, 12 Nov 2003 14:22:58 -0400 (AST) Received: (qmail 9143 invoked from network); 12 Nov 2003 18:22:58 -0000 Received: from unknown (HELO ?10.0.2.5?) (216.208.117.7) by 205.178.180.9 with SMTP; 12 Nov 2003 18:22:58 -0000 Subject: Re: performance optimzations From: Rod Taylor To: Suchandra Thapa Cc: Postgresql Performance In-Reply-To: <1068658430.1071.42.camel@hepcat> References: <1068658430.1071.42.camel@hepcat> Content-Type: text/plain Message-Id: <1068661380.30452.17.camel@jester> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Wed, 12 Nov 2003 13:23:01 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/73 X-Sequence-Number: 4633 On Wed, 2003-11-12 at 12:34, Suchandra Thapa wrote: > I'm moving a webmail service over to use a postgresql database for > storage and wanted to get any tips for optimizing performance. The > machine will be a multiprocessor (either 2 or 4 cpu ) system with a raid > array. What layout should be used? I was thinking using about using a > raid 1+0 array to hold the database but since I can use different array > types, would it be better to use 1+0 for the wal logs and a raid 5 for > the database? How much in total storage? If you have (or will have) > ~6 disks, go for RAID 5 otherwise 10 is probably appropriate. > The database gets fairly heavy activity (the system handles about 500MB > of incoming and about 750MB of outgoing emails daily). I have a fairly > free rein in regards to the system's layout as well as how the > applications will interact with the database since I'm writing the > code. These are archived permanently -- ~450GB of annual data? Or is the data removed upon delivery? From pgsql-performance-owner@postgresql.org Wed Nov 12 14:28:37 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 47BA8D1B535 for ; Wed, 12 Nov 2003 18:28:37 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 40876-03 for ; Wed, 12 Nov 2003 14:28:07 -0400 (AST) Received: from shire.ontko.com (shire.ontko.com [199.164.165.1]) by svr1.postgresql.org (Postfix) with ESMTP id 40BCFD1B522 for ; Wed, 12 Nov 2003 14:28:06 -0400 (AST) Received: from nick (bilbo.ontko.com [199.164.165.101]) by shire.ontko.com (8.12.3/8.12.3/Debian-6.6) with SMTP id hACIS45f017833; Wed, 12 Nov 2003 13:28:05 -0500 Reply-To: From: "Nick Fankhauser" To: "Tom Lane" Cc: "Manfred Koizar" , "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long Date: Wed, 12 Nov 2003 13:27:53 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) In-Reply-To: <10144.1068657035@sss.pgh.pa.us> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/74 X-Sequence-Number: 4634 > You might have to resort to brute force, like "set enable_nestloop=false". > Just out of curiosity, what do you get if you do that? I get a different plan, but similar execution time: Limit (cost=323437.13..323437.13 rows=1 width=115) (actual time=170921.89..170922.95 rows=1000 loops=1) -> Sort (cost=323437.13..323437.13 rows=1 width=115) (actual time=170921.89..170922.26 rows=1001 loops=1) Sort Key: max((actor.actor_full_name)::text), count(case_data.case_id) -> Aggregate (cost=323437.08..323437.12 rows=1 width=115) (actual time=170849.94..170898.06 rows=3457 loops=1) -> Group (cost=323437.08..323437.09 rows=3 width=115) (actual time=170849.90..170873.60 rows=5880 loops=1) -> Sort (cost=323437.08..323437.08 rows=3 width=115) (actual time=170847.97..170850.21 rows=5880 loops=1) Sort Key: actor.actor_id -> Hash Join (cost=253333.29..323437.06 rows=3 width=115) (actual time=122873.80..170814.27 rows=5880 loops=1) Hash Cond: ("outer".case_id = "inner".case_id) -> Seq Scan on case_data (cost=0.00..60368.16 rows=1947116 width=39) (actual time=12.95..43542.25 rows=1947377 loops=1) -> Hash (cost=253333.28..253333.28 rows=3 width=76) (actual time=122844.40..122844.40 rows=0 loops=1) -> Hash Join (cost=6.02..253333.28 rows=3 width=76) (actual time=24992.70..122810.32 rows=5883 loops=1) Hash Cond: ("outer".actor_id = "inner".actor_id) -> Seq Scan on actor_case_assignment (cost=0.00..209980.49 rows=8669349 width=34) (actual time=9.13..85504.05 rows=8670467 loops=1) -> Hash (cost=6.01..6.01 rows=1 width=42) (actual time=24926.56..24926.56 rows=0 loops=1) -> Index Scan using actor_full_name_uppercase on actor (cost=0.00..6.01 rows=1 width=42) (actual time=51.67..24900.53 rows=3502 loops=1) Index Cond: ((actor_full_name_uppercase >= 'SANDERS'::character varying) AND (actor_full_name_uppercase < 'SANDERT'::character varying)) Filter: (actor_full_name_uppercase ~~ 'SANDERS%'::text) Total runtime: 170925.93 msec (19 rows) -Nick From pgsql-performance-owner@postgresql.org Wed Nov 12 14:55:45 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9EF9AD1B53F for ; Wed, 12 Nov 2003 18:55:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 40919-09 for ; Wed, 12 Nov 2003 14:55:13 -0400 (AST) Received: from node-4024d1a2.mdw.onnet.us.uu.net (nsit-s229-45.uchicago.edu [128.135.229.45]) by svr1.postgresql.org (Postfix) with ESMTP id C6ABDD1B502 for ; Wed, 12 Nov 2003 14:55:04 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by node-4024d1a2.mdw.onnet.us.uu.net (Postfix) with ESMTP id AC60D37D0A; Wed, 12 Nov 2003 12:51:10 -0600 (CST) Subject: Re: performance optimzations From: Suchandra Thapa To: Rod Taylor Cc: Postgresql Performance In-Reply-To: <1068661380.30452.17.camel@jester> References: <1068658430.1071.42.camel@hepcat> <1068661380.30452.17.camel@jester> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-RPIt6DKUfon9QoHE8LNb" Organization: Message-Id: <1068663070.5193.21.camel@hepcat> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.2.2 (1.2.2-5) Date: 12 Nov 2003 12:51:10 -0600 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/75 X-Sequence-Number: 4635 --=-RPIt6DKUfon9QoHE8LNb Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Wed, 2003-11-12 at 12:23, Rod Taylor wrote: > On Wed, 2003-11-12 at 12:34, Suchandra Thapa wrote: > > I'm moving a webmail service over to use a postgresql database for > > storage and wanted to get any tips for optimizing performance. The > > machine will be a multiprocessor (either 2 or 4 cpu ) system with a raid > > array. What layout should be used? I was thinking using about using a > > raid 1+0 array to hold the database but since I can use different array > > types, would it be better to use 1+0 for the wal logs and a raid 5 for > > the database? >=20 > How much in total storage? If you have (or will have) > ~6 disks, go > for RAID 5 otherwise 10 is probably appropriate. I'm not sure but I believe there are about 6-8 10K scsi drives on the system. There is quite a bit of storage to spare currently so I think=20 > > The database gets fairly heavy activity (the system handles about 500MB > > of incoming and about 750MB of outgoing emails daily). I have a fairly > > free rein in regards to the system's layout as well as how the > > applications will interact with the database since I'm writing the > > code. >=20 > These are archived permanently -- ~450GB of annual data? Or is the data > removed upon delivery? No, it's more like hotmail. Some users may keep mail for a longer term but a lot of the mail probably gets deleted fairly quickly. The database load will be mixed with a insertions due to deliveries, queries by the webmail system, and deletions from pop and webmail. --=20 Suchandra Thapa --=-RPIt6DKUfon9QoHE8LNb Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQA/soEe6nShCjt5AZIRAt21AKCFSDhmJ7O/Q/vm1UM3+b9PKM4XggCeOI7K naNwR0tmY5gOldNONquIAzc= =WvjL -----END PGP SIGNATURE----- --=-RPIt6DKUfon9QoHE8LNb-- From pgsql-performance-owner@postgresql.org Tue Nov 11 16:17:38 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D8916D1B55F for ; Tue, 11 Nov 2003 20:17:36 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 21535-02 for ; Tue, 11 Nov 2003 16:17:07 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id 46CC9D1B4EB for ; Tue, 11 Nov 2003 16:17:05 -0400 (AST) Received: (qmail 581 invoked from network); 11 Nov 2003 20:18:01 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 11 Nov 2003 20:18:01 -0000 Message-ID: <3FB29532.9060107@trade-india.com> Date: Thu, 13 Nov 2003 01:46:50 +0530 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Josh Berkus Cc: Christopher Browne , pgsql-performance@postgresql.org Subject: Re: Suggestions for benchmarking 7.4RC2 against 7.3 References: <3FB247FE.4080808@trade-india.com> <3FB24C15.1020907@trade-india.com> <200311111025.28806.josh@agliodbs.com> In-Reply-To: <200311111025.28806.josh@agliodbs.com> Content-Type: multipart/alternative; boundary="------------070000020205060807000709" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/47 X-Sequence-Number: 4607 This is a multi-part message in MIME format. --------------070000020205060807000709 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Josh Berkus wrote: >Rajesh, Chris, > > > >>I got the osdb benchmark running last week, and had to separate client >>from server. I had to jump through a fair number of hoops including >>copying data files over to the server. The benchmark software needs a >>bit more work... >> >> > >What about the OSDL's TPC-derivative benchmarks? That's a much more >respected database test, and probably less buggy than OSDB. > > > Hmm... really sorry! my pg_dump | psql is almost finishing in next 20 mins. creating indexes at the moment :) Really sorry can't rollback and delay anymore becoz my website is *unavailable* for past 30 mins. I ran OSDB .15 version and pg_bench . Regds Mallah. --------------070000020205060807000709 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Josh Berkus wrote:
Rajesh, Chris,

  
I got the osdb benchmark running last week, and had to separate client
from server.  I had to jump through a fair number of hoops including
copying data files over to the server.  The benchmark software needs a
bit more work...
    

What about the OSDL's TPC-derivative benchmarks?   That's a much more 
respected database test, and probably less buggy than OSDB.
Hmm... really sorry! my
pg_dump | psql is almost finishing in next 20 mins.

creating indexes at the moment :)

Really sorry can't rollback and delay anymore becoz my
website is *unavailable* for past 30 mins.

I ran OSDB .15 version  and pg_bench .


Regds
Mallah.




--------------070000020205060807000709-- From pgsql-performance-owner@postgresql.org Tue Nov 11 17:25:39 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id BA74BD1B531 for ; Tue, 11 Nov 2003 21:25:38 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 32185-02 for ; Tue, 11 Nov 2003 17:25:09 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id 0AFB0D1B4FE for ; Tue, 11 Nov 2003 17:25:07 -0400 (AST) Received: (qmail 4307 invoked from network); 11 Nov 2003 21:26:18 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 11 Nov 2003 21:26:18 -0000 Message-ID: <3FB2A540.2000808@trade-india.com> Date: Thu, 13 Nov 2003 02:55:20 +0530 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Christopher Browne Cc: pgsql-performance@postgresql.org Subject: Re: Suggestions for benchmarking 7.4RC2 against 7.3 References: <3FB247FE.4080808@trade-india.com> <3FB24C15.1020907@trade-india.com> In-Reply-To: Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/48 X-Sequence-Number: 4608

RC2 is running in production without any apparent problems
till now.  Well its difficult to say at the moment how much speed
gain is there unless the heavy duty batch SQL scripts are run by
cron.

Count(*) and group by on large tables are significantly (5x) faster
and better error reporting has made it easier to spot the faulty data.
eg in fkey violation.

Will post the OSDB .15 versions' results on 7.3 & 7.4 soon.

Regds
Mallah.

Christopher Browne wrote:
After a long battle with technology,mallah@trade-india.com (Rajesh Kumar Mallah), an earthling, wrote:
  
the error mentioned in first email has been overcome
by running osdb on the same machine hosting the DB server.
    

Yes, it seems unrealistic to try to run the "client" on a separate
host from the database.  

I got the osdb benchmark running last week, and had to separate client
from server.  I had to jump through a fair number of hoops including
copying data files over to the server.  The benchmark software needs a
bit more work...

From pgsql-performance-owner@postgresql.org Wed Nov 12 20:23:08 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B1B05D1B50B for ; Wed, 12 Nov 2003 22:28:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78209-10 for ; Wed, 12 Nov 2003 18:27:43 -0400 (AST) Received: from email06.aon.at (WARSL402PIP3.highway.telekom.at [195.3.96.75]) by svr1.postgresql.org (Postfix) with SMTP id 2DD90D1B4EB for ; Wed, 12 Nov 2003 18:27:24 -0400 (AST) Received: (qmail 453032 invoked from network); 12 Nov 2003 22:27:01 -0000 Received: from m170p002.dipool.highway.telekom.at (HELO cantor) ([62.46.11.34]) (envelope-sender ) by qmail6rs.highway.telekom.at (qmail-ldap-1.03) with SMTP for ; 12 Nov 2003 22:27:01 -0000 From: Manfred Koizar To: Cc: "Tom Lane" , "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long Date: Wed, 12 Nov 2003 23:25:54 +0100 Message-ID: References: <10144.1068657035@sss.pgh.pa.us> In-Reply-To: X-Mailer: Forte Agent 1.93/32.576 English (American) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/78 X-Sequence-Number: 4638 On Wed, 12 Nov 2003 13:27:53 -0500, "Nick Fankhauser" wrote: > >> You might have to resort to brute force, like "set enable_nestloop=false". > -> Seq Scan on >actor_case_assignment (cost=0.00..209980.49 rows=8669349 width=34) (actual >time=9.13..85504.05 rows=8670467 loops=1) Does actor_case_assignment contain more columns than just the two ids? If yes, do these additional fields account for ca. 70 bytes per tuple? If not, try VACUUM FULL ANALYSE actor_case_assignment; > -> Index Scan using >actor_full_name_uppercase on actor (cost=0.00..6.01 rows=1 width=42) >(actual time=51.67..24900.53 rows=3502 loops=1) This same index scan on actor has been much faster in your previous postings (677ms, 3200ms), probably due to caching effects. 7ms per tuple returned looks like a lot of disk seeks are involved. Is clustering actor on actor_full_name_uppercase an option or would this slow down other queries? Servus Manfred From pgsql-performance-owner@postgresql.org Wed Nov 12 20:23:05 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D392DD1B4ED for ; Wed, 12 Nov 2003 22:29:59 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 81605-07 for ; Wed, 12 Nov 2003 18:29:31 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id 29DCED1B50C for ; Wed, 12 Nov 2003 18:29:29 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id B52BF1FC7; Wed, 12 Nov 2003 17:29:30 -0500 (EST) To: Suchandra Thapa Cc: pgsql-performance@postgresql.org Subject: Re: performance optimzations From: Neil Conway In-Reply-To: <1068658430.1071.42.camel@hepcat> (Suchandra Thapa's message of "12 Nov 2003 11:34:41 -0600") References: <1068658430.1071.42.camel@hepcat> Date: Wed, 12 Nov 2003 17:29:30 -0500 Message-ID: <87oevh9q5x.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/77 X-Sequence-Number: 4637 Suchandra Thapa writes: > I was thinking using about using a raid 1+0 array to hold the > database but since I can use different array types, would it be > better to use 1+0 for the wal logs and a raid 5 for the database? It has been recommended on this list that getting a RAID controller with a battery-backed cache is pretty essential to getting good performance. Search the list archives for lots more discussion about RAID configurations. -Neil From pgsql-performance-owner@postgresql.org Wed Nov 12 20:23:26 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9FE74D1B557 for ; Thu, 13 Nov 2003 00:20:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 05831-01 for ; Wed, 12 Nov 2003 20:19:34 -0400 (AST) Received: from node-4024d1a2.mdw.onnet.us.uu.net (nsit-s227-189.uchicago.edu [128.135.227.189]) by svr1.postgresql.org (Postfix) with ESMTP id 975DBD1B553 for ; Wed, 12 Nov 2003 20:19:30 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by node-4024d1a2.mdw.onnet.us.uu.net (Postfix) with ESMTP id D878E37D0A; Wed, 12 Nov 2003 18:15:58 -0600 (CST) Subject: Re: performance optimzations From: Suchandra Thapa To: Neil Conway Cc: pgsql-performance@postgresql.org In-Reply-To: <87oevh9q5x.fsf@mailbox.samurai.com> References: <1068658430.1071.42.camel@hepcat> <87oevh9q5x.fsf@mailbox.samurai.com> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-nUmlQp/VCc50mmNOJhgO" Organization: Message-Id: <1068682558.5193.47.camel@hepcat> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.2.2 (1.2.2-5) Date: 12 Nov 2003 18:15:58 -0600 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/79 X-Sequence-Number: 4639 --=-nUmlQp/VCc50mmNOJhgO Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Wed, 2003-11-12 at 16:29, Neil Conway wrote: > Suchandra Thapa writes: > > I was thinking using about using a raid 1+0 array to hold the > > database but since I can use different array types, would it be > > better to use 1+0 for the wal logs and a raid 5 for the database? >=20 > It has been recommended on this list that getting a RAID controller > with a battery-backed cache is pretty essential to getting good > performance. Search the list archives for lots more discussion about > RAID configurations. The server is already using a raid controller with battery backed ram and the cache set to write back (the server is on a ups so power failures shouldn't cause problems). I'll look at the list archives for RAID information.=20=20 --=20 Suchandra Thapa --=-nUmlQp/VCc50mmNOJhgO Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQA/ss0+6nShCjt5AZIRAj9GAJ9PQ3+CH11zo2zvQ3EXkD0pcE++ngCfSE8O +TjrbFQ0jBgOZmwkLFUVEIw= =5ZuI -----END PGP SIGNATURE----- --=-nUmlQp/VCc50mmNOJhgO-- From pgsql-performance-owner@postgresql.org Thu Nov 13 00:07:12 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id AA988D1B527 for ; Thu, 13 Nov 2003 04:07:04 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 41081-06 for ; Thu, 13 Nov 2003 00:06:32 -0400 (AST) Received: from houston.familyhealth.com.au (fhnet.arach.net.au [203.22.197.21]) by svr1.postgresql.org (Postfix) with ESMTP id 8E5D4D1B4FE for ; Thu, 13 Nov 2003 00:06:31 -0400 (AST) Received: from familyhealth.com.au (work-40.internal [192.168.0.40]) by houston.familyhealth.com.au (8.12.9p1/8.12.9) with ESMTP id hAD46UoD004575 for ; Thu, 13 Nov 2003 12:06:30 +0800 (WST) (envelope-from chriskl@familyhealth.com.au) Message-ID: <3FB303E9.20903@familyhealth.com.au> Date: Thu, 13 Nov 2003 12:09:13 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: PostgreSQL Performance Subject: Query question Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/80 X-Sequence-Number: 4640 Hi, I have coded some improvements to phpPgAdmin that I think are pretty cool. Basicaly, once you are browsing the results of an arbitrary SELECT query, you can still sort by columns, regardless of the underlying ORDER BY of the SELECT. I do this like this: SELECT * FROM (arbitrary subquery) AS sub ORDER BY 1,3; Now, this all works fine, but I want to know if this is efficient or not. Does doing a select of a select cause serious performance degradation? Chris From pgsql-performance-owner@postgresql.org Thu Nov 13 00:36:28 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 400BDD1B527 for ; Thu, 13 Nov 2003 04:36:22 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 44534-05 for ; Thu, 13 Nov 2003 00:35:51 -0400 (AST) Received: from mail.hive.nj2.inquent.com (mc.carriermail.com [205.178.180.9]) by svr1.postgresql.org (Postfix) with SMTP id 4C23DD1B535 for ; Thu, 13 Nov 2003 00:35:35 -0400 (AST) Received: (qmail 27827 invoked from network); 13 Nov 2003 04:34:32 -0000 Received: from unknown (HELO ?192.168.1.199?) (134.22.68.129) by 205.178.180.9 with SMTP; 13 Nov 2003 04:34:32 -0000 Subject: Re: performance optimzations From: Rod Taylor To: Suchandra Thapa Cc: Postgresql Performance In-Reply-To: <1068663070.5193.21.camel@hepcat> References: <1068658430.1071.42.camel@hepcat> <1068661380.30452.17.camel@jester> <1068663070.5193.21.camel@hepcat> Content-Type: text/plain Message-Id: <1068698132.31569.46.camel@jester> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Wed, 12 Nov 2003 23:35:33 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/81 X-Sequence-Number: 4641 > > How much in total storage? If you have (or will have) > ~6 disks, go > > for RAID 5 otherwise 10 is probably appropriate. > > I'm not sure but I believe there are about 6-8 10K scsi drives on the > system. There is quite a bit of storage to spare currently so I think I see.. With 8 drives, you'll probably want to go with RAID 5. It grows beyond that point fairly well with a decent controller card. Be sure to have some battery backed write cache on the raid card (128MB goes a long way). > > > The database gets fairly heavy activity (the system handles about 500MB > > > of incoming and about 750MB of outgoing emails daily). I have a fairly > No, it's more like hotmail. Some users may keep mail for a longer term > but a lot of the mail probably gets deleted fairly quickly. The > database load will be mixed with a insertions due to deliveries, queries > by the webmail system, and deletions from pop and webmail. You might consider having the mailserver gzip the emails prior to injection into the database (turn off compression in PostgreSQL) and decompress the data on the webserver for display to the client. Now you have about 7 times the number of emails in memory. It's easier to toss a webserver at the problem than make the database bigger in size. Take the savings in CPU on the DB and add it to ram. 1200MB of compressed mail is about 200MB? Assume email descriptive material (subject, from, etc.), account structure, indexes... so about 400MB for one days worth of information? You may want to consider keeping the compressed email in a separate table than the information describing it. It would mean descriptive information is more likely to be in RAM, where the body probably doesn't matter as much (you view them 1 at a time, subjects tend to be listed all at once). Most clients will be interested in say the last 7 days worth of data? Great.. Start out with 4GB ram on a good Dual CPU -- Opterons seem to work quite well -- and make sure the motherboard can hold double that in memory for an upgrade sometime next year when you've become popular. I firmly believe lots of RAM is the answer to most IO issues until you start getting into large sets of active data (>50GB). 64GB ram is fairly cheap compared to ongoing maintenance of the 30+ drive system required to get decent throughput. From pgsql-performance-owner@postgresql.org Thu Nov 13 01:21:25 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B02B0D1B4E1 for ; Thu, 13 Nov 2003 05:21:22 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 54620-09 for ; Thu, 13 Nov 2003 01:20:52 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id B6A91D1B507 for ; Thu, 13 Nov 2003 01:20:51 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3908779; Wed, 12 Nov 2003 21:21:35 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: Christopher Kings-Lynne , PostgreSQL Performance Subject: Re: Query question Date: Wed, 12 Nov 2003 21:20:28 -0800 User-Agent: KMail/1.4.3 References: <3FB303E9.20903@familyhealth.com.au> In-Reply-To: <3FB303E9.20903@familyhealth.com.au> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311122120.28882.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/82 X-Sequence-Number: 4642 Chris, > SELECT * FROM (arbitrary subquery) AS sub ORDER BY 1,3; > > Now, this all works fine, but I want to know if this is efficient or not. > > Does doing a select of a select cause serious performance degradation? It would be better if you could strip out the inner sort, but I can understand why that might not be possible in all cases. The only thing you're adding to the query is a second SORT step, so it shouldn't require any more time/memory than the query's first SORT did. -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Thu Nov 13 02:03:36 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id AA9ACD1B50B for ; Thu, 13 Nov 2003 06:03:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 66527-08 for ; Thu, 13 Nov 2003 02:03:04 -0400 (AST) Received: from node-4024d1a2.mdw.onnet.us.uu.net (nsit-s227-70.uchicago.edu [128.135.227.70]) by svr1.postgresql.org (Postfix) with ESMTP id C2DE8D1B535 for ; Thu, 13 Nov 2003 02:02:46 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by node-4024d1a2.mdw.onnet.us.uu.net (Postfix) with ESMTP id 1E96A37D7E; Wed, 12 Nov 2003 23:58:45 -0600 (CST) Subject: Re: performance optimzations From: Suchandra Thapa To: Rod Taylor Cc: Postgresql Performance In-Reply-To: <1068698132.31569.46.camel@jester> References: <1068658430.1071.42.camel@hepcat> <1068661380.30452.17.camel@jester> <1068663070.5193.21.camel@hepcat> <1068698132.31569.46.camel@jester> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-BcogzFHj1OVm6xatBeRz" Organization: Message-Id: <1068703124.12978.27.camel@hepcat> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.2.2 (1.2.2-5) Date: 12 Nov 2003 23:58:45 -0600 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/83 X-Sequence-Number: 4643 --=-BcogzFHj1OVm6xatBeRz Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Wed, 2003-11-12 at 22:35, Rod Taylor wrote: > You may want to consider keeping the compressed email in a separate > table than the information describing it. It would mean descriptive > information is more likely to be in RAM, where the body probably doesn't > matter as much (you view them 1 at a time, subjects tend to be listed > all at once). Thanks for the suggestions. Splitting the load between several machines was the original intent of moving the storage from the file system to a database. I believe the schema I'm already using splits out the body due to the size of some attachments. Luckily the code already gzips the email body and abbreviates common email headers so storing compressed emails isn't a problem.=20=20 > Most clients will be interested in say the last 7 days worth of data?=20 > Great.. Start out with 4GB ram on a good Dual CPU -- Opterons seem to > work quite well -- and make sure the motherboard can hold double that in > memory for an upgrade sometime next year when you've become popular. Unfortunately, the hardware available is pretty much fixed in regards to the system. I can play around with the raid configurations and have some limited choice in regards to the raid controller and number of drivers but that's about all in terms of hardware.=20=20=20=20 > I firmly believe lots of RAM is the answer to most IO issues until you > start getting into large sets of active data (>50GB). 64GB ram is fairly > cheap compared to ongoing maintenance of the 30+ drive system required > to get decent throughput. The current file system holding the user and email information indicates the current data has about 64GB (70K accounts, I'm not sure how many are active but 50% might be good guess). This seems to be somewhat of a steady state however. --=20 Suchandra Thapa --=-BcogzFHj1OVm6xatBeRz Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQA/sx2U6nShCjt5AZIRAmVjAKCGAeAFlKcKMrY7sULbwNwoiy59rQCfcqki Ui3n6yqsdei2GQCOzAQOi9o= =q+x/ -----END PGP SIGNATURE----- --=-BcogzFHj1OVm6xatBeRz-- From pgsql-performance-owner@postgresql.org Thu Nov 13 03:51:26 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2DB48D1B502 for ; Thu, 13 Nov 2003 07:51:25 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 86098-08 for ; Thu, 13 Nov 2003 03:50:55 -0400 (AST) Received: from rtlocal.trade-india.com (unknown [61.16.154.82]) by svr1.postgresql.org (Postfix) with SMTP id B24A4D1B51C for ; Thu, 13 Nov 2003 03:50:47 -0400 (AST) Received: (qmail 6741 invoked from network); 13 Nov 2003 08:09:08 -0000 Received: from unknown (HELO trade-india.com) (unknown) by unknown with SMTP; 13 Nov 2003 08:09:08 -0000 From: Rajesh Kumar Mallah Organization: Infocom Network Limited To: pgsql-performance@postgresql.org Subject: IN surpasses NOT EXISTS in 7.4RC2 ?? Date: Thu, 13 Nov 2003 13:23:39 +0530 User-Agent: KMail/1.5.1 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311131323.39762.mallah@trade-india.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/84 X-Sequence-Number: 4644 Hi, NOT EXISTS is taking almost double time than NOT IN . I know IN has been optimised in 7.4 but is anything wrong with the NOT EXISTS? I have vaccumed , analyze and run the query many times still not in is faster than exists :> Regds Mallah. NOT IN PLAN tradein_clients=# explain analyze SELECT count(*) from general.profile_master where profile_id not in (select profile_id from general.account_profiles ) ; QUERY PLAN ------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=32238.19..32238.19 rows=1 width=0) (actual time=5329.206..5329.207 rows=1 loops=1) -> Seq Scan on profile_master (cost=4458.25..31340.38 rows=359125 width=0) (actual time=1055.496..4637.908 rows=470386 loops=1) Filter: (NOT (hashed subplan)) SubPlan -> Seq Scan on account_profiles (cost=0.00..3817.80 rows=256180 width=4) (actual time=0.061..507.811 rows=256180 loops=1) Total runtime: 5337.591 ms (6 rows) tradein_clients=# explain analyze SELECT count(*) from general.profile_master where not exists (select profile_id from general.account_profiles where profile_id=general.profile_master.profile_id ) ; QUERY PLAN --------------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=1674981.97..1674981.97 rows=1 width=0) (actual time=14600.386..14600.387 rows=1 loops=1) -> Seq Scan on profile_master (cost=0.00..1674084.16 rows=359125 width=0) (actual time=13.687..13815.798 rows=470386 loops=1) Filter: (NOT (subplan)) SubPlan -> Index Scan using account_profiles_profile_id on account_profiles (cost=0.00..4.59 rows=2 width=4) (actual time=0.013..0.013 rows=0 loops=718250) Index Cond: (profile_id = $0) Total runtime: 14600.531 ms From pgsql-performance-owner@postgresql.org Thu Nov 13 09:48:09 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5F3ADD1B535 for ; Thu, 13 Nov 2003 13:48:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 41251-01 for ; Thu, 13 Nov 2003 09:47:39 -0400 (AST) Received: from mail.hive.nj2.inquent.com (mc.carriermail.com [205.178.180.9]) by svr1.postgresql.org (Postfix) with SMTP id D78DAD1B51E for ; Thu, 13 Nov 2003 09:47:29 -0400 (AST) Received: (qmail 10716 invoked from network); 13 Nov 2003 13:47:27 -0000 Received: from unknown (HELO ?192.168.1.199?) (134.22.68.129) by 205.178.180.9 with SMTP; 13 Nov 2003 13:47:27 -0000 Subject: Re: performance optimzations From: Rod Taylor To: Suchandra Thapa Cc: Postgresql Performance In-Reply-To: <1068703124.12978.27.camel@hepcat> References: <1068658430.1071.42.camel@hepcat> <1068661380.30452.17.camel@jester> <1068663070.5193.21.camel@hepcat> <1068698132.31569.46.camel@jester> <1068703124.12978.27.camel@hepcat> Content-Type: text/plain Message-Id: <1068731245.31569.152.camel@jester> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Thu, 13 Nov 2003 08:47:33 -0500 Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/85 X-Sequence-Number: 4645 > > Most clients will be interested in say the last 7 days worth of data? > > Great.. Start out with 4GB ram on a good Dual CPU -- Opterons seem to > > work quite well -- and make sure the motherboard can hold double that in > > memory for an upgrade sometime next year when you've become popular. > > Unfortunately, the hardware available is pretty much fixed in regards to > the system. I can play around with the raid configurations and have > some limited choice in regards to the raid controller and number of > drivers but that's about all in terms of hardware. Good luck then. Unless the configuration takes into account incremental additions in ram and disk, sustained growth could get very expensive. I guess that depends on the business plan expectations. This just puts more emphasis to offload everything you can onto machines that can multiply. > The current file system holding the user and email information indicates > the current data has about 64GB (70K accounts, I'm not sure how many are > active but 50% might be good guess). This seems to be somewhat of a > steady state however. 35k clients checking their mail daily isn't so bad. Around 10 pages per second peak load? From pgsql-performance-owner@postgresql.org Thu Nov 13 12:25:01 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1EC20D1B51C for ; Thu, 13 Nov 2003 14:57:54 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 51627-04 for ; Thu, 13 Nov 2003 10:57:26 -0400 (AST) Received: from candle.pha.pa.us (unknown [207.106.42.251]) by svr1.postgresql.org (Postfix) with ESMTP id D6FF1D1B515 for ; Thu, 13 Nov 2003 10:57:20 -0400 (AST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.11.6) id hADEuv409079; Thu, 13 Nov 2003 09:56:57 -0500 (EST) From: Bruce Momjian Message-Id: <200311131456.hADEuv409079@candle.pha.pa.us> Subject: Re: IN surpasses NOT EXISTS in 7.4RC2 ?? In-Reply-To: <200311131323.39762.mallah@trade-india.com> To: Rajesh Kumar Mallah Date: Thu, 13 Nov 2003 09:56:57 -0500 (EST) Cc: pgsql-performance@postgresql.org X-Mailer: ELM [version 2.4ME+ PL108 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/88 X-Sequence-Number: 4648 Rajesh Kumar Mallah wrote: > > Hi, > > NOT EXISTS is taking almost double time than NOT IN . > I know IN has been optimised in 7.4 but is anything > wrong with the NOT EXISTS? > > I have vaccumed , analyze and run the query many times > still not in is faster than exists :> Seems fine. In 7.4, NOT IN will often be faster that NOT EXISTS. NOT EXISTS didn't change --- there are restrictions on how far we can optimize NOT EXISTS. NOT IN has just become much faster in 7.4. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073 From pgsql-performance-owner@postgresql.org Thu Nov 13 11:51:11 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A1F40D1B527 for ; Thu, 13 Nov 2003 15:20:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 53496-08 for ; Thu, 13 Nov 2003 11:19:51 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 63DA9D1B4FE for ; Thu, 13 Nov 2003 11:19:47 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hADFJp19026332; Thu, 13 Nov 2003 10:19:52 -0500 (EST) To: Rajesh Kumar Mallah Cc: pgsql-performance@postgresql.org Subject: Re: IN surpasses NOT EXISTS in 7.4RC2 ?? In-reply-to: <200311131323.39762.mallah@trade-india.com> References: <200311131323.39762.mallah@trade-india.com> Comments: In-reply-to Rajesh Kumar Mallah message dated "Thu, 13 Nov 2003 13:23:39 +0530" Date: Thu, 13 Nov 2003 10:19:51 -0500 Message-ID: <26331.1068736791@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/86 X-Sequence-Number: 4646 Rajesh Kumar Mallah writes: > NOT EXISTS is taking almost double time than NOT IN . > I know IN has been optimised in 7.4 but is anything > wrong with the NOT EXISTS? That's the expected behavior in 7.4. EXISTS in the style you are using it effectively forces a nestloop-with-inner-indexscan implementation. As of 7.4, IN can do that, but it can do several other things too, including the hash-type plan you have here. So assuming that the planner chooses the right plan choice (not always a given ;-)) IN should be as fast or faster than EXISTS in all cases. regards, tom lane From pgsql-performance-owner@postgresql.org Thu Nov 13 12:39:07 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id BE1D9D1B50E for ; Thu, 13 Nov 2003 15:32:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 51621-10 for ; Thu, 13 Nov 2003 11:31:44 -0400 (AST) Received: from bramble.mmrd.com (unknown [65.217.53.66]) by svr1.postgresql.org (Postfix) with ESMTP id 3FE00D1B4F1 for ; Thu, 13 Nov 2003 11:31:39 -0400 (AST) Received: from thorn.mmrd.com (thorn.mmrd.com [172.25.10.100]) by bramble.mmrd.com (8.12.8/8.12.8) with ESMTP id hADEafcM008710; Thu, 13 Nov 2003 09:36:41 -0500 Received: from gnvex001.mmrd.com (gnvex001.mmrd.com [192.168.3.55]) by thorn.mmrd.com (8.11.6/8.11.6) with ESMTP id hADFVgl26052; Thu, 13 Nov 2003 10:31:43 -0500 Received: from camel.mmrd.com ([172.25.5.213]) by gnvex001.mmrd.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2657.72) id V3JRTPVA; Thu, 13 Nov 2003 10:31:41 -0500 Subject: Re: IN surpasses NOT EXISTS in 7.4RC2 ?? From: Robert Treat To: Rajesh Kumar Mallah Cc: pgsql-performance@postgresql.org In-Reply-To: <200311131323.39762.mallah@trade-india.com> References: <200311131323.39762.mallah@trade-india.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.8 Date: 13 Nov 2003 10:31:42 -0500 Message-Id: <1068737502.10946.18338.camel@camel> Mime-Version: 1.0 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/90 X-Sequence-Number: 4650 It is believed that the IN optimization can lead to faster IN times than EXIST times on some queries, the extent of which is still a bit of an unknown. (Incidentally is there an FAQ item on this that needs updating?) Does the not exist query produce worse results in 7.4 than it did in 7.3? Robert Treat On Thu, 2003-11-13 at 02:53, Rajesh Kumar Mallah wrote: > > Hi, > > NOT EXISTS is taking almost double time than NOT IN . > I know IN has been optimised in 7.4 but is anything > wrong with the NOT EXISTS? > > I have vaccumed , analyze and run the query many times > still not in is faster than exists :> > > > Regds > Mallah. > > NOT IN PLAN > > tradein_clients=# explain analyze SELECT count(*) from general.profile_master where > profile_id not in (select profile_id from general.account_profiles ) ; > QUERY PLAN > ------------------------------------------------------------------------------------------------------------------------------------- > Aggregate (cost=32238.19..32238.19 rows=1 width=0) (actual time=5329.206..5329.207 rows=1 loops=1) > -> Seq Scan on profile_master (cost=4458.25..31340.38 rows=359125 width=0) (actual time=1055.496..4637.908 rows=470386 loops=1) > Filter: (NOT (hashed subplan)) > SubPlan > -> Seq Scan on account_profiles (cost=0.00..3817.80 rows=256180 width=4) (actual time=0.061..507.811 rows=256180 loops=1) > Total runtime: 5337.591 ms > (6 rows) > > > tradein_clients=# explain analyze SELECT count(*) from general.profile_master where not exists > (select profile_id from general.account_profiles where profile_id=general.profile_master.profile_id ) ; > > QUERY PLAN > --------------------------------------------------------------------------------------------------------------------------------------------------------------- > Aggregate (cost=1674981.97..1674981.97 rows=1 width=0) (actual time=14600.386..14600.387 rows=1 loops=1) > -> Seq Scan on profile_master (cost=0.00..1674084.16 rows=359125 width=0) (actual time=13.687..13815.798 rows=470386 loops=1) > Filter: (NOT (subplan)) > SubPlan > -> Index Scan using account_profiles_profile_id on account_profiles (cost=0.00..4.59 rows=2 width=4) (actual time=0.013..0.013 rows=0 loops=718250) > Index Cond: (profile_id = $0) > Total runtime: 14600.531 ms > > > > ---------------------------(end of broadcast)--------------------------- > TIP 2: you can get off all lists at once with the unregister command > (send "unregister YourEmailAddressHere" to majordomo@postgresql.org) -- Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL From pgsql-performance-owner@postgresql.org Thu Nov 13 12:00:52 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5B4A9D1B50E for ; Thu, 13 Nov 2003 15:52:58 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 63479-03 for ; Thu, 13 Nov 2003 11:52:31 -0400 (AST) Received: from bramble.mmrd.com (unknown [65.217.53.66]) by svr1.postgresql.org (Postfix) with ESMTP id 5995ED1B527 for ; Thu, 13 Nov 2003 11:52:26 -0400 (AST) Received: from thorn.mmrd.com (thorn.mmrd.com [172.25.10.100]) by bramble.mmrd.com (8.12.8/8.12.8) with ESMTP id hADEvOcM008935; Thu, 13 Nov 2003 09:57:24 -0500 Received: from gnvex001.mmrd.com (gnvex001.mmrd.com [192.168.3.55]) by thorn.mmrd.com (8.11.6/8.11.6) with ESMTP id hADFqPl26573; Thu, 13 Nov 2003 10:52:25 -0500 Received: from camel.mmrd.com ([172.25.5.213]) by gnvex001.mmrd.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2657.72) id V3JRTPZH; Thu, 13 Nov 2003 10:52:24 -0500 Subject: Re: Value of Quad vs. Dual Processor machine From: Robert Treat To: Jeff Cc: Chris Field , ron.l.johnson@cox.net, pgsql-performance@postgresql.org In-Reply-To: <20031112092807.14b18bf4.threshar@torgo.978.org> References: <1068593567.2063.8.camel@tuxsrva> <1068600291.30969.53.camel@haggis.homelan> <001401c3a8c2$8c8fec30$040210ac@win2000> <20031112092807.14b18bf4.threshar@torgo.978.org> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.8 Date: 13 Nov 2003 10:52:25 -0500 Message-Id: <1068738745.10945.18380.camel@camel> Mime-Version: 1.0 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/87 X-Sequence-Number: 4647 On Wed, 2003-11-12 at 09:28, Jeff wrote: > On Tue, 11 Nov 2003 21:13:19 -0500 > "Chris Field" wrote: > > > we are looking at Xeon, We are currently running it on a quad sun v880 > > compiled to be 64bit and have been getting dreadful performance. I > > don't think we really have much to gain from going 64bit. > > > > > By chance, are you running 7.3.4 on that sun? > If so, try this: > export CFLAGS=-02 > ./configure > > and rebuild PG. > > Before 7.4 PG was build with _no_ optimization on Solaris. > Recompiling gives __HUGE__ (notice the underscores) performance gains. > > And onto the dual vs quad. > > PG will only use 1 cpu / connection / query. > > So if your machine iwll have 1-2 queries running at a time those other 2 > proc's will sit around idling. However if you are going to have a bunch > going, 4 cpus will be most useful. One of hte nicest things to do for > PG is more ram and fast IO. It really loves those things. > We've just started kicking around the idea of moving one of our boxes to a quad-proc machine from a dual. Under normal circumstances the 2 processors handle maybe 200 transactions per second with 90% system idle. However we have people who occasionally run historical reports on our data, and those reports are fairly CPU intensive. Usually it is not a problem for the main web system, but when pg_dump is running, that is also cpu intensive, so we end up with two highly cpu intensive items running on our machine, and we start to notice issues on the main web system. Robert Treat -- Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL From pgsql-performance-owner@postgresql.org Thu Nov 13 13:00:51 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 69D98D1B502 for ; Thu, 13 Nov 2003 17:00:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78207-07 for ; Thu, 13 Nov 2003 13:00:12 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 497B2D1B531 for ; Thu, 13 Nov 2003 13:00:11 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hADH0B19027152; Thu, 13 Nov 2003 12:00:11 -0500 (EST) To: Robert Treat Cc: Rajesh Kumar Mallah , pgsql-performance@postgresql.org Subject: Re: IN surpasses NOT EXISTS in 7.4RC2 ?? In-reply-to: <1068737502.10946.18338.camel@camel> References: <200311131323.39762.mallah@trade-india.com> <1068737502.10946.18338.camel@camel> Comments: In-reply-to Robert Treat message dated "13 Nov 2003 10:31:42 -0500" Date: Thu, 13 Nov 2003 12:00:10 -0500 Message-ID: <27151.1068742810@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/91 X-Sequence-Number: 4651 Robert Treat writes: > Does the not exist query produce worse results in 7.4 than it did in > 7.3? EXISTS should work the same as before. regards, tom lane From pgsql-performance-owner@postgresql.org Thu Nov 13 13:20:17 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 67303D1B54B for ; Thu, 13 Nov 2003 17:20:15 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78988-07 for ; Thu, 13 Nov 2003 13:19:44 -0400 (AST) Received: from bramble.mmrd.com (unknown [65.217.53.66]) by svr1.postgresql.org (Postfix) with ESMTP id 3E382D1B558 for ; Thu, 13 Nov 2003 13:19:43 -0400 (AST) Received: from thorn.mmrd.com (thorn.mmrd.com [172.25.10.100]) by bramble.mmrd.com (8.12.8/8.12.8) with ESMTP id hADGOgcM009669; Thu, 13 Nov 2003 11:24:42 -0500 Received: from gnvex001.mmrd.com (gnvex001.mmrd.com [192.168.3.55]) by thorn.mmrd.com (8.11.6/8.11.6) with ESMTP id hADHJgl28561; Thu, 13 Nov 2003 12:19:42 -0500 Received: from camel.mmrd.com ([172.25.5.213]) by gnvex001.mmrd.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2657.72) id V3JRTQL2; Thu, 13 Nov 2003 12:19:41 -0500 Subject: Re: IN surpasses NOT EXISTS in 7.4RC2 ?? From: Robert Treat To: Tom Lane Cc: Rajesh Kumar Mallah , pgsql-performance@postgresql.org In-Reply-To: <27151.1068742810@sss.pgh.pa.us> References: <200311131323.39762.mallah@trade-india.com> <1068737502.10946.18338.camel@camel> <27151.1068742810@sss.pgh.pa.us> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.8 Date: 13 Nov 2003 12:19:42 -0500 Message-Id: <1068743982.10945.18510.camel@camel> Mime-Version: 1.0 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/92 X-Sequence-Number: 4652 On Thu, 2003-11-13 at 12:00, Tom Lane wrote: > Robert Treat writes: > > Does the not exist query produce worse results in 7.4 than it did in > > 7.3? > > EXISTS should work the same as before. > right. the original poster is asking if there is "something wrong with exist" based on the comparison to IN, he needs to compare it vs. 7.3 EXISTS to determine if something is wrong. Robert Treat -- Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL From pgsql-performance-owner@postgresql.org Thu Nov 13 14:14:53 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 73AEDD1B508 for ; Thu, 13 Nov 2003 18:14:52 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 95024-06 for ; Thu, 13 Nov 2003 14:14:22 -0400 (AST) Received: from mail.libertyrms.com (unknown [209.167.124.227]) by svr1.postgresql.org (Postfix) with ESMTP id 7AD38D1B4FE for ; Thu, 13 Nov 2003 14:14:20 -0400 (AST) Received: from [10.1.2.130] (helo=dba2) by mail.libertyrms.com with esmtp (Exim 3.22 #3 (Debian)) id 1AKLyx-0007uk-00 for ; Thu, 13 Nov 2003 13:14:15 -0500 Received: by dba2 (Postfix, from userid 1019) id 9DE12CB49; Thu, 13 Nov 2003 13:14:15 -0500 (EST) Date: Thu, 13 Nov 2003 13:14:15 -0500 From: Andrew Sullivan To: pgsql-performance@postgresql.org Subject: strange estimate for number of rows Message-ID: <20031113181415.GB25546@libertyrms.info> Mail-Followup-To: Andrew Sullivan , pgsql-performance@postgresql.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/93 X-Sequence-Number: 4653 Hi all, I've one here that I cannot fathom. Any suggestions? We have a table, call it tablename, where we're selecting by a range of dates and an identifier. (This is redacted, obviously): \d tablename Column | Type | Modifiers --------------------+--------------------------+-------------------- id | integer | not null transaction_date | timestamp with time zone | not null product_id | integer | not null Indexes: "trans_posted_trans_date_idx" btree (transaction_date, product_id) The statistics on transaction_date and product_id are set to 1000. Everything is all analysed nicely. But I'm getting a poor plan, because of an estimate that the number of rows to be returned is about double how many actually are: explain analyse select * from transactions_posted where transaction_date >= '2003-9-1' and transaction_date < '2003-10-1' and product_id = 2; QUERY PLAN --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Seq Scan on transactions_posted (cost=0.00..376630.33 rows=700923 width=91) (actual time=8422.253..36176.078 rows=316029 loops=1) Filter: ((transaction_date >= '2003-09-01 00:00:00-04'::timestamp with time zone) AND (transaction_date < '2003-10-01 00:00:00-04'::timestamp with time zone) AND (product_id = 2)) Total runtime: 36357.630 ms (3 rows) SET enable_seqscan = off; explain analyse select * from transactions_posted where transaction_date >= '2003-9-1' and transaction_date < '2003-10-1' and product_id = 2; QUERY PLAN ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Index Scan using trans_posted_trans_date_idx on transactions_posted (cost=0.00..1088862.56 rows=700923 width=91) (actual time=35.214..14816.257 rows=316029 loops=1) Index Cond: ((transaction_date >= '2003-09-01 00:00:00-04'::timestamp with time zone) AND (transaction_date < '2003-10-01 00:00:00-04'::timestamp with time zone) AND (product_id = 2)) Total runtime: 15009.816 ms (3 rows) SELECT attname,null_frac,avg_width,n_distinct,correlation FROM pg_stats where tablename = 'transactions_posted' AND attname in ('transaction_date','product_id'); attname | null_frac | avg_width | n_distinct | correlation ------------------+-----------+-----------+------------+------------- product_id | 0 | 4 | 2 | 0.200956 transaction_date | 0 | 8 | -0.200791 | 0.289248 Any ideas? I'm loathe to recommend cluster, since the data will not stay clustered. A -- ---- Andrew Sullivan 204-4141 Yonge Street Afilias Canada Toronto, Ontario Canada M2P 2A8 +1 416 646 3304 x110 From pgsql-performance-owner@postgresql.org Thu Nov 13 14:28:49 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8EF5DD1B4EB for ; Thu, 13 Nov 2003 18:28:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 98312-08 for ; Thu, 13 Nov 2003 14:28:17 -0400 (AST) Received: from award.gotdns.org (pcp01097251pcs.tsclos01.al.comcast.net [68.62.129.152]) by svr1.postgresql.org (Postfix) with ESMTP id 2A7CFD1B508 for ; Thu, 13 Nov 2003 14:28:16 -0400 (AST) Received: (qmail 9067 invoked by uid 81); 13 Nov 2003 18:28:15 -0000 Received: from 68.62.129.152 (SquirrelMail authenticated user award) by award.gotdns.org with HTTP; Thu, 13 Nov 2003 12:28:15 -0600 (CST) Message-ID: <23123.68.62.129.152.1068748095.squirrel@award.gotdns.org> Date: Thu, 13 Nov 2003 12:28:15 -0600 (CST) Subject: Union+group by planner estimates way off? From: "Arthur Ward" To: pgsql-performance@postgresql.org User-Agent: SquirrelMail/1.4.1 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/94 X-Sequence-Number: 4654 On 7.4 RC2, I'm seeing a case where the query planner estimates are way out of line after grouping the result of a union. I've tried adjusting the statistics targets up to 200, and it made no difference in the planner's estimates. The point of the full query this came from is that it also has an aggregate function that produces a space-delimited list of commodity & fak for each id. Does anyone have any suggestions on tweaks to apply or ways to rewrite this? Is this one of those ugly corners where the query planner doesn't have a clue how to estimate this (seeing the nice round 200 estimate makes me suspicious)? EXPLAIN ANALYZE SELECT id FROM (SELECT id, commodity FROM commodities WHERE commodity IS NOT NULL UNION SELECT id, fak FROM commodities WHERE fak IS NOT NULL ) all_commodities GROUP BY id; QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------------------------------------------- HashAggregate (cost=15939.16..15939.16 rows=200 width=4) (actual time=3537.281..3680.418 rows=83306 loops=1) -> Subquery Scan all_commodities (cost=14002.00..15697.02 rows=96858 width=4) (actual time=2268.052..3214.996 rows=95715 loops=1) -> Unique (cost=14002.00..14728.44 rows=96858 width=15) (actual time=2268.043..2881.688 rows=95715 loops=1) -> Sort (cost=14002.00..14244.15 rows=96858 width=15) (actual time=2268.037..2527.083 rows=100008 loops=1) Sort Key: id, commodity -> Append (cost=0.00..5034.42 rows=96858 width=15) (actual time=7.402..1220.320 rows=100008 loops=1) -> Subquery Scan "*SELECT* 1" (cost=0.00..2401.23 rows=36831 width=15) (actual time=7.398..590.004 rows=39772 loops=1) -> Seq Scan on commodities (cost=0.00..2032.92 rows=36831 width=15) (actual time=7.388..468.415 rows=39772 loops=1) Filter: (commodity IS NOT NULL) -> Subquery Scan "*SELECT* 2" (cost=0.00..2633.19 rows=60027 width=14) (actual time=0.016..408.160 rows=60236 loops=1) -> Seq Scan on commodities (cost=0.00..2032.92 rows=60027 width=14) (actual time=0.010..221.635 rows=60236 loops=1) Filter: (fak IS NOT NULL) Total runtime: 3783.009 ms (13 rows) From pgsql-performance-owner@postgresql.org Thu Nov 13 14:47:23 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5F095D1B508 for ; Thu, 13 Nov 2003 18:47:22 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 04220-10 for ; Thu, 13 Nov 2003 14:46:52 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 5E127D1B554 for ; Thu, 13 Nov 2003 14:46:51 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hADIkp19005807; Thu, 13 Nov 2003 13:46:51 -0500 (EST) To: "Arthur Ward" Cc: pgsql-performance@postgresql.org Subject: Re: Union+group by planner estimates way off? In-reply-to: <23123.68.62.129.152.1068748095.squirrel@award.gotdns.org> References: <23123.68.62.129.152.1068748095.squirrel@award.gotdns.org> Comments: In-reply-to "Arthur Ward" message dated "Thu, 13 Nov 2003 12:28:15 -0600" Date: Thu, 13 Nov 2003 13:46:51 -0500 Message-ID: <5806.1068749211@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/95 X-Sequence-Number: 4655 "Arthur Ward" writes: > EXPLAIN ANALYZE SELECT id FROM > (SELECT id, commodity FROM commodities WHERE commodity IS NOT NULL > UNION > SELECT id, fak FROM commodities WHERE fak IS NOT NULL > ) all_commodities GROUP BY id; > QUERY > PLAN > ---------------------------------------------------------------------------------------------------------------------------------------------------------------- > HashAggregate (cost=15939.16..15939.16 rows=200 width=4) (actual > time=3537.281..3680.418 rows=83306 loops=1) > -> Subquery Scan all_commodities (cost=14002.00..15697.02 rows=96858 > width=4) (actual time=2268.052..3214.996 rows=95715 loops=1) It's falling back to a default estimate because it doesn't know how to find any statistics for the output of a sub-select. I have a TODO somewhere about burrowing down into sub-selects to see if the output maps directly to a column that we'd have stats for ... but it's not done yet. In this particular case the inaccurate estimate doesn't matter too much, I think, although it might be encouraging the system to select hash aggregation since it thinks the hashtable will be pretty small. If the estimate were getting used to plan higher-up plan steps then it could be a bigger problem. regards, tom lane From pgsql-performance-owner@postgresql.org Thu Nov 13 14:57:00 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EC188D1B522 for ; Thu, 13 Nov 2003 18:56:58 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 05355-09 for ; Thu, 13 Nov 2003 14:56:28 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id E78D4D1B50E for ; Thu, 13 Nov 2003 14:56:26 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hADIuR19005883; Thu, 13 Nov 2003 13:56:27 -0500 (EST) To: Andrew Sullivan Cc: pgsql-performance@postgresql.org Subject: Re: strange estimate for number of rows In-reply-to: <20031113181415.GB25546@libertyrms.info> References: <20031113181415.GB25546@libertyrms.info> Comments: In-reply-to Andrew Sullivan message dated "Thu, 13 Nov 2003 13:14:15 -0500" Date: Thu, 13 Nov 2003 13:56:26 -0500 Message-ID: <5882.1068749786@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/96 X-Sequence-Number: 4656 Andrew Sullivan writes: > The statistics on transaction_date and product_id are set to 1000. > Everything is all analysed nicely. But I'm getting a poor plan, > because of an estimate that the number of rows to be returned is > about double how many actually are: > explain analyse select * from transactions_posted where > transaction_date >= '2003-9-1' and transaction_date < '2003-10-1' and > product_id = 2; Are the estimates accurate for queries on the two columns individually, ie ... where transaction_date >= '2003-9-1' and transaction_date < '2003-10-1' ... where product_id = 2 If so, the problem is that there's a correlation between transaction_date and product_id, which the system cannot model because it has no multi-column statistics. However, given that the estimate is only off by about a factor of 2, you'd still be getting the wrong plan even if the estimate were perfect, because the estimated costs differ by nearly a factor of 3. Given the actual runtimes, I'm thinking maybe you want to reduce random_page_cost. What are you using for that now? regards, tom lane From pgsql-performance-owner@postgresql.org Thu Nov 13 15:19:42 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6B048D1B522 for ; Thu, 13 Nov 2003 19:19:39 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 16409-01 for ; Thu, 13 Nov 2003 15:19:08 -0400 (AST) Received: from award.gotdns.org (pcp01097251pcs.tsclos01.al.comcast.net [68.62.129.152]) by svr1.postgresql.org (Postfix) with ESMTP id 059F4D1B508 for ; Thu, 13 Nov 2003 15:19:07 -0400 (AST) Received: (qmail 9850 invoked by uid 81); 13 Nov 2003 19:19:07 -0000 Received: from 68.62.129.152 (SquirrelMail authenticated user award) by award.gotdns.org with HTTP; Thu, 13 Nov 2003 13:19:07 -0600 (CST) Message-ID: <50572.68.62.129.152.1068751147.squirrel@award.gotdns.org> In-Reply-To: <5806.1068749211@sss.pgh.pa.us> References: <23123.68.62.129.152.1068748095.squirrel@award.gotdns.org> <5806.1068749211@sss.pgh.pa.us> Date: Thu, 13 Nov 2003 13:19:07 -0600 (CST) Subject: Re: Union+group by planner estimates way off? From: "Arthur Ward" To: "Tom Lane" Cc: pgsql-performance@postgresql.org User-Agent: SquirrelMail/1.4.1 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/97 X-Sequence-Number: 4657 > In this particular case the inaccurate estimate doesn't matter too much, > I think, although it might be encouraging the system to select hash > aggregation since it thinks the hashtable will be pretty small. If the > estimate were getting used to plan higher-up plan steps then it could > be a bigger problem. That's my problem; this is a subselect feeding in to a larger query. That wrong estimate causes the planner to select a nested-loop at the next step up. At 83,000 rows, the word is "ouch!" At any rate, I discovered this while dissecting a giant & slow query. Hence, while disabling nested-loop joins avoids this particular pitfall, it's not good for the bigger picture. I think I'm going to end up splitting this larger query into smaller parts and reassemble the pieces in the application so I can push some smarts past other subselect boundaries. For my purposes, that should skirt the issue of union+group estimates not being calculated. As always, thanks for the fast answers! From pgsql-performance-owner@postgresql.org Thu Nov 13 15:43:01 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 54F97D1B4E1 for ; Thu, 13 Nov 2003 19:43:01 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 16175-08 for ; Thu, 13 Nov 2003 15:42:30 -0400 (AST) Received: from www.postgresql.com (www.postgresql.com [200.46.204.209]) by svr1.postgresql.org (Postfix) with ESMTP id 69341D1B522 for ; Thu, 13 Nov 2003 15:42:29 -0400 (AST) Received: from mail.libertyrms.com (unknown [209.167.124.227]) by www.postgresql.com (Postfix) with ESMTP id A5679CF48D4 for ; Thu, 13 Nov 2003 15:42:28 -0400 (AST) Received: from dev7.int.libertyrms.com ([10.1.2.225] helo=libertyrms.info) by mail.libertyrms.com with esmtp (Exim 3.22 #3 (Debian)) id 1AKNHK-0000hW-00; Thu, 13 Nov 2003 14:37:18 -0500 Message-ID: <3FB3DD1E.2000509@libertyrms.info> Date: Thu, 13 Nov 2003 14:35:58 -0500 From: Daniel Manley User-Agent: Mozilla/5.0 (X11; U; Linux i686; ru-RU; rv:1.5) Gecko/20031016 X-Accept-Language: en, fr, en-us MIME-Version: 1.0 To: Tom Lane Cc: pgsql-performance@postgresql.org Subject: Re: strange estimate for number of rows References: <20031113181415.GB25546@libertyrms.info> <5882.1068749786@sss.pgh.pa.us> In-Reply-To: <5882.1068749786@sss.pgh.pa.us> Content-Type: text/plain; charset=KOI8-R; format=flowed Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/100 X-Sequence-Number: 4660 Hi, I'm the lead developer on the project this concerns (forgive my newbiness on this list). We tried a couple of scenarios with effective_cache_size=60000, cpu_index_tuple_cost=0.0001 and random_page_cost=2 with no change in the plan. explain analyse select * from tablename where transaction_date >= '2003-9-1' and transaction_date < '2003-10-1'; ----------------------------------------------------------------------------------------------------------------------------------- Seq Scan on tablename (cost=0.00..348199.14 rows=1180724 width=91) (actual time=7727.668..36286.898 rows=579238 loops=1) Filter: ((transaction_date >= '2003-09-01 00:00:00+00'::timestamp with time zone) AND (transaction_date < '2003-10-01 00:00:00+00'::timestamp with time zone)) Total runtime: 36625.351 ms explain analyse select * from transactions_posted where product_id = 2; ----------------------------------------------------------------------------------------------------------------------------------- Seq Scan on transactions_posted (cost=0.00..319767.95 rows=6785237 width=91) (actual time=0.091..35596.328 rows=5713877 loops=1) Filter: (product_id = 2) Total runtime: 38685.373 ms The product_id alone gives a difference of a millions rows from estimate to actual, vs. the factor of 2 from the transaction_date. Dan Manley Tom Lane �����: >Andrew Sullivan writes: > > >>The statistics on transaction_date and product_id are set to 1000. >>Everything is all analysed nicely. But I'm getting a poor plan, >>because of an estimate that the number of rows to be returned is >>about double how many actually are: >> >> > > > >>explain analyse select * from transactions_posted where >>transaction_date >= '2003-9-1' and transaction_date < '2003-10-1' and >>product_id = 2; >> >> > >Are the estimates accurate for queries on the two columns individually, >ie >... where transaction_date >= '2003-9-1' and transaction_date < '2003-10-1' >... where product_id = 2 > >If so, the problem is that there's a correlation between >transaction_date and product_id, which the system cannot model because >it has no multi-column statistics. > >However, given that the estimate is only off by about a factor of 2, >you'd still be getting the wrong plan even if the estimate were perfect, >because the estimated costs differ by nearly a factor of 3. > >Given the actual runtimes, I'm thinking maybe you want to reduce >random_page_cost. What are you using for that now? > > regards, tom lane > >---------------------------(end of broadcast)--------------------------- >TIP 4: Don't 'kill -9' the postmaster > > From pgsql-performance-owner@postgresql.org Thu Nov 13 15:39:25 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2F6A3D1B540 for ; Thu, 13 Nov 2003 19:39:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 15957-06 for ; Thu, 13 Nov 2003 15:38:53 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id EB0A5D1B4F1 for ; Thu, 13 Nov 2003 15:38:51 -0400 (AST) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3914758 for pgsql-performance@postgresql.org; Thu, 13 Nov 2003 11:39:21 -0800 Content-Type: text/plain; charset="us-ascii" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: pgsql-performance@postgresql.org Subject: Storage space, RAM for NUMERIC Date: Thu, 13 Nov 2003 11:36:05 -0800 User-Agent: KMail/1.4.3 MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200311131136.05066.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/99 X-Sequence-Number: 4659 Folks, How would I calculate storage space/required ram on a 50-digit NUMERIC? And the docs state that NUMERIC is slow. Is this just slow for calculatio= ns=20 (due to the conversion to float & back) or slow for index lookups as well? --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Thu Nov 13 16:19:40 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C6E11D1B559 for ; Thu, 13 Nov 2003 20:19:39 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 19869-05 for ; Thu, 13 Nov 2003 16:19:10 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id A20CED1B522 for ; Thu, 13 Nov 2003 16:19:08 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hADKJ819011161; Thu, 13 Nov 2003 15:19:09 -0500 (EST) To: Daniel Manley Cc: pgsql-performance@postgresql.org Subject: Re: strange estimate for number of rows In-reply-to: <3FB3DD1E.2000509@libertyrms.info> References: <20031113181415.GB25546@libertyrms.info> <5882.1068749786@sss.pgh.pa.us> <3FB3DD1E.2000509@libertyrms.info> Comments: In-reply-to Daniel Manley message dated "Thu, 13 Nov 2003 14:35:58 -0500" Date: Thu, 13 Nov 2003 15:19:08 -0500 Message-ID: <11160.1068754748@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/101 X-Sequence-Number: 4661 Daniel Manley writes: > The product_id alone gives a difference of a millions rows from estimate > to actual, vs. the factor of 2 from the transaction_date. You should be thinking in terms of ratios, not absolute difference. The rows estimate for product_id doesn't look too bad to me; the one for transaction_date is much further off (a factor of 2). Which is odd, because the system can usually do all right on range estimates if you've let it run an ANALYZE with enough histogram bins. Could we see the pg_stats row for transaction_date? > We tried a couple of scenarios with effective_cache_size=60000, > cpu_index_tuple_cost=0.0001 and random_page_cost=2 with no change in the > plan. Since you need about a factor of 3 change in the cost estimate to get it to switch plans, changing random_page_cost by a factor of 2 ain't gonna do it (the other two numbers are second-order adjustments unlikely to have much effect, I think). Try 1.5, or even less ... of course, you have to keep an eye on your other queries and make sure they don't go nuts, but from what I've heard about your hardware setup a low random_page_cost isn't out of line with the physical realities. regards, tom lane From pgsql-performance-owner@postgresql.org Thu Nov 13 17:37:39 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 85FAFD1B4ED for ; Thu, 13 Nov 2003 21:37:32 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38680-04 for ; Thu, 13 Nov 2003 17:37:03 -0400 (AST) Received: from mail.libertyrms.com (unknown [209.167.124.227]) by svr1.postgresql.org (Postfix) with ESMTP id 5D12FD1B535 for ; Thu, 13 Nov 2003 17:37:01 -0400 (AST) Received: from [10.1.2.130] (helo=dba2) by mail.libertyrms.com with esmtp (Exim 3.22 #3 (Debian)) id 1AKP9D-0002cd-00 for ; Thu, 13 Nov 2003 16:37:03 -0500 Received: by dba2 (Postfix, from userid 1019) id 4C0FFCB49; Thu, 13 Nov 2003 16:37:03 -0500 (EST) Date: Thu, 13 Nov 2003 16:37:03 -0500 From: Andrew Sullivan To: pgsql-performance@postgresql.org Subject: Re: strange estimate for number of rows Message-ID: <20031113213703.GJ25546@libertyrms.info> Mail-Followup-To: Andrew Sullivan , pgsql-performance@postgresql.org References: <20031113181415.GB25546@libertyrms.info> <5882.1068749786@sss.pgh.pa.us> <3FB3DD1E.2000509@libertyrms.info> <11160.1068754748@sss.pgh.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <11160.1068754748@sss.pgh.pa.us> User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/102 X-Sequence-Number: 4662 On Thu, Nov 13, 2003 at 03:19:08PM -0500, Tom Lane wrote: > because the system can usually do all right on range estimates if you've > let it run an ANALYZE with enough histogram bins. Could we see the > pg_stats row for transaction_date? Do you want the whole thing? I left out the really verbose bits when I posted this in the original: SELECT attname,null_frac,avg_width,n_distinct,correlation FROM pg_stats where tablename = 'transactions_posted' AND attname in ('transaction_date','product_id'); attname | null_frac | avg_width | n_distinct | correlation ------------------+-----------+-----------+------------+------------- product_id | 0 | 4 | 2 | 0.200956 transaction_date | 0 | 8 | -0.200791 | 0.289248 > > Since you need about a factor of 3 change in the cost estimate to get it to > switch plans, changing random_page_cost by a factor of 2 ain't gonna do > it (the other two numbers are second-order adjustments unlikely to have > much effect, I think). Try 1.5, or even less ... of course, you have to > keep an eye on your other queries and make sure they don't go nuts, but > from what I've heard about your hardware setup a low random_page_cost > isn't out of line with the physical realities. Actually, this one's on an internal box, and I think 1.5 is too low -- it's really just a pair of mirrored SCSI disks on a PCI controller in this case. That does the trick, though, so maybe I'm just being too conservantive. A -- ---- Andrew Sullivan 204-4141 Yonge Street Afilias Canada Toronto, Ontario Canada M2P 2A8 +1 416 646 3304 x110 From pgsql-performance-owner@postgresql.org Thu Nov 13 17:53:23 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 45187D1B540 for ; Thu, 13 Nov 2003 21:53:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38400-08 for ; Thu, 13 Nov 2003 17:52:52 -0400 (AST) Received: from mail.libertyrms.com (unknown [209.167.124.227]) by svr1.postgresql.org (Postfix) with ESMTP id D34AAD1B531 for ; Thu, 13 Nov 2003 17:52:49 -0400 (AST) Received: from [10.1.2.130] (helo=dba2) by mail.libertyrms.com with esmtp (Exim 3.22 #3 (Debian)) id 1AKPOW-0002sK-00 for ; Thu, 13 Nov 2003 16:52:52 -0500 Received: by dba2 (Postfix, from userid 1019) id 037E9CB49; Thu, 13 Nov 2003 16:52:51 -0500 (EST) Date: Thu, 13 Nov 2003 16:52:51 -0500 From: Andrew Sullivan To: pgsql-performance@postgresql.org Subject: Re: strange estimate for number of rows Message-ID: <20031113215251.GL25546@libertyrms.info> Mail-Followup-To: Andrew Sullivan , pgsql-performance@postgresql.org References: <20031113181415.GB25546@libertyrms.info> <5882.1068749786@sss.pgh.pa.us> <3FB3DD1E.2000509@libertyrms.info> <11160.1068754748@sss.pgh.pa.us> <20031113213703.GJ25546@libertyrms.info> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20031113213703.GJ25546@libertyrms.info> User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/103 X-Sequence-Number: 4663 On Thu, Nov 13, 2003 at 04:37:03PM -0500, Andrew Sullivan wrote: > Actually, this one's on an internal box, and I think 1.5 is too low > -- it's really just a pair of mirrored SCSI disks on a PCI controller > in this case. That does the trick, though, so maybe I'm just being > too conservantive. I spoke too soon. I'd left enable_seqscan=off set. It doesn't actually prefer an indexscan until I set the random_page_cost to .5. I think that's a little unrealistic ;-) A -- ---- Andrew Sullivan 204-4141 Yonge Street Afilias Canada Toronto, Ontario Canada M2P 2A8 +1 416 646 3304 x110 From pgsql-performance-owner@postgresql.org Fri Nov 14 03:19:22 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 7A387D1B4EB for ; Fri, 14 Nov 2003 07:19:20 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 30604-05 for ; Fri, 14 Nov 2003 03:18:50 -0400 (AST) Received: from rtlocal.trade-india.com (unknown [61.16.154.82]) by svr1.postgresql.org (Postfix) with SMTP id 56930D1B4F1 for ; Fri, 14 Nov 2003 03:18:46 -0400 (AST) Received: (qmail 12618 invoked from network); 14 Nov 2003 07:37:06 -0000 Received: from unknown (HELO trade-india.com) (unknown) by unknown with SMTP; 14 Nov 2003 07:37:06 -0000 From: Rajesh Kumar Mallah Organization: Infocom Network Limited To: pgsql-performance@postgresql.org Subject: Help with count(*) Date: Fri, 14 Nov 2003 12:51:38 +0530 User-Agent: KMail/1.5.1 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311141251.38786.mallah@trade-india.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/104 X-Sequence-Number: 4664 Hi , my database seems to be taking too long for a select count(*) i think there are lot of dead rows. I do a vacuum full it improves bu again the performance drops in a short while , can anyone please tell me if anything worng with my fsm settings current fsm=55099264 (not sure how i calculated it) Regds Mallah tradein_clients=# SELECT count(*) from data_bank.profiles ; +--------+ | count | +--------+ | 123065 | +--------+ (1 row) Time: 49756.969 ms tradein_clients=# tradein_clients=# tradein_clients=# VACUUM full verbose analyze data_bank.profiles ; INFO: vacuuming "data_bank.profiles" INFO: "profiles": found 0 removable, 369195 nonremovable row versions in 43423 pages DETAIL: 246130 dead row versions cannot be removed yet. Nonremovable row versions range from 136 to 2036 bytes long. There were 427579 unused item pointers. Total free space (including removable row versions) is 178536020 bytes. 15934 pages are or will become empty, including 0 at the end of the table. 38112 pages containing 178196624 free bytes are potential move destinations. CPU 1.51s/0.63u sec elapsed 23.52 sec. INFO: index "profiles_pincode" now contains 369195 row versions in 3353 pages DETAIL: 0 index row versions were removed. 379 index pages have been deleted, 379 are currently reusable. CPU 0.20s/0.24u sec elapsed 22.73 sec. INFO: index "profiles_city" now contains 369195 row versions in 3411 pages DETAIL: 0 index row versions were removed. 1030 index pages have been deleted, 1030 are currently reusable. CPU 0.17s/0.21u sec elapsed 20.67 sec. INFO: index "profiles_branch" now contains 369195 row versions in 2209 pages DETAIL: 0 index row versions were removed. 783 index pages have been deleted, 783 are currently reusable. CPU 0.07s/0.14u sec elapsed 6.38 sec. INFO: index "profiles_area_code" now contains 369195 row versions in 2606 pages DETAIL: 0 index row versions were removed. 856 index pages have been deleted, 856 are currently reusable. CPU 0.11s/0.17u sec elapsed 19.62 sec. INFO: index "profiles_source" now contains 369195 row versions in 3137 pages DETAIL: 0 index row versions were removed. 1199 index pages have been deleted, 1199 are currently reusable. CPU 0.14s/0.12u sec elapsed 9.95 sec. INFO: index "co_name_index_idx" now contains 368742 row versions in 3945 pages DETAIL: 0 index row versions were removed. 0 index pages have been deleted, 0 are currently reusable. CPU 0.19s/0.69u sec elapsed 11.56 sec. INFO: index "address_index_idx" now contains 368898 row versions in 4828 pages DETAIL: 0 index row versions were removed. 0 index pages have been deleted, 0 are currently reusable. CPU 0.16s/0.61u sec elapsed 9.17 sec. INFO: index "profiles_exp_cat" now contains 153954 row versions in 2168 pages DETAIL: 0 index row versions were removed. 0 index pages have been deleted, 0 are currently reusable. CPU 0.07s/0.25u sec elapsed 3.14 sec. INFO: index "profiles_imp_cat" now contains 73476 row versions in 1030 pages DETAIL: 0 index row versions were removed. 0 index pages have been deleted, 0 are currently reusable. CPU 0.05s/0.11u sec elapsed 8.73 sec. INFO: index "profiles_manu_cat" now contains 86534 row versions in 1193 pages DETAIL: 0 index row versions were removed. 0 index pages have been deleted, 0 are currently reusable. CPU 0.03s/0.13u sec elapsed 1.44 sec. INFO: index "profiles_serv_cat" now contains 19256 row versions in 267 pages DETAIL: 0 index row versions were removed. 0 index pages have been deleted, 0 are currently reusable. CPU 0.01s/0.01u sec elapsed 0.25 sec. INFO: index "profiles_pid" now contains 369195 row versions in 812 pages DETAIL: 0 index row versions were removed. 0 index pages have been deleted, 0 are currently reusable. CPU 0.03s/0.12u sec elapsed 0.41 sec. INFO: index "profiles_pending_branch_id" now contains 0 row versions in 1 pages DETAIL: 0 index row versions were removed. 0 index pages have been deleted, 0 are currently reusable. CPU 0.00s/0.00u sec elapsed 0.00 sec. INFO: "profiles": moved 0 row versions, truncated 43423 to 43423 pages DETAIL: CPU 1.76s/3.01u sec elapsed 60.39 sec. INFO: vacuuming "pg_toast.pg_toast_39873340" INFO: "pg_toast_39873340": found 0 removable, 65 nonremovable row versions in 15 pages DETAIL: 0 dead row versions cannot be removed yet. Nonremovable row versions range from 47 to 2034 bytes long. There were 0 unused item pointers. Total free space (including removable row versions) is 17672 bytes. 0 pages are or will become empty, including 0 at the end of the table. 14 pages containing 17636 free bytes are potential move destinations. CPU 0.00s/0.00u sec elapsed 0.77 sec. INFO: index "pg_toast_39873340_index" now contains 65 row versions in 2 pages DETAIL: 0 index pages have been deleted, 0 are currently reusable. CPU 0.00s/0.00u sec elapsed 0.46 sec. INFO: "pg_toast_39873340": moved 0 row versions, truncated 15 to 15 pages DETAIL: CPU 0.00s/0.00u sec elapsed 0.00 sec. INFO: analyzing "data_bank.profiles" INFO: "profiles": 43423 pages, 123065 rows sampled, 123065 estimated total rows VACUUM Time: 246989.138 ms tradein_clients=# SELECT count(*) from data_bank.profiles ; +--------+ | count | +--------+ | 123065 | +--------+ (1 row) Time: 4978.725 ms tradein_clients=# IMPORVED but still not very good. Regds Mallah. From pgsql-performance-owner@postgresql.org Fri Nov 14 03:44:51 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C3BFDD1B535 for ; Fri, 14 Nov 2003 07:44:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38130-02 for ; Fri, 14 Nov 2003 03:44:18 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 7747BD1B559 for ; Fri, 14 Nov 2003 03:44:00 -0400 (AST) Received: from daithan.intranet.pspl.co.in shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Fri, 14 Nov 2003 00:44:02 -0700 From: Shridhar Daithankar To: pgsql-performance@postgresql.org Subject: Re: Help with count(*) Date: Fri, 14 Nov 2003 13:00:34 +0530 User-Agent: KMail/1.5.4 References: <200311141251.38786.mallah@trade-india.com> In-Reply-To: <200311141251.38786.mallah@trade-india.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Message-Id: <200311141300.34397.shridhar_daithankar@myrealbox.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/105 X-Sequence-Number: 4665 On Friday 14 November 2003 12:51, Rajesh Kumar Mallah wrote: > Hi , > > my database seems to be taking too long for a select count(*) > i think there are lot of dead rows. I do a vacuum full it improves > bu again the performance drops in a short while , > can anyone please tell me if anything worng with my fsm settings > current fsm=3D55099264 (not sure how i calculated it) If you don't need exact count, you can use statistics. Just analyze frequen= tly=20 and you will get the statistics. and I didn't exact;y understand this in the text. INFO: =A0"profiles": found 0 removable, 369195 nonremovable row versions in= =20 43423 pages DETAIL: =A0246130 dead row versions cannot be removed yet. Is there a transaction holoding up large amount of stuff? Shridhar From pgsql-performance-owner@postgresql.org Fri Nov 14 05:45:39 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 18E07D1B54B for ; Fri, 14 Nov 2003 09:45:38 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 55482-03 for ; Fri, 14 Nov 2003 05:45:09 -0400 (AST) Received: from ALPHA9.ITS.MONASH.EDU.AU (alpha9.its.monash.edu.au [130.194.1.9]) by svr1.postgresql.org (Postfix) with ESMTP id 01DD4D1B544 for ; Fri, 14 Nov 2003 05:45:07 -0400 (AST) Received: from localhost ([130.194.13.81]) by vaxh.its.monash.edu.au (PMDF V5.2-31 #39306) with ESMTP id <01L319WO9YN28WZH8O@vaxh.its.monash.edu.au> for pgsql-performance@postgresql.org; Fri, 14 Nov 2003 20:38:33 +1100 Received: from kapow.its.monash.edu.au (localhost.its.monash.edu.au [127.0.0.1]) by localhost (Postfix) with ESMTP id 5B6EF1B0002 for ; Fri, 14 Nov 2003 20:38:33 +1100 (EST) Received: from bruce.csse.monash.edu.au (bruce.csse.monash.edu.au [130.194.64.3]) by kapow.its.monash.edu.au (Postfix) with ESMTP id 45563124003 for ; Fri, 14 Nov 2003 20:38:33 +1100 (EST) Received: from bruce.csse.monash.edu.au (localhost [127.0.0.1]) by bruce.csse.monash.edu.au (8.12.8+Sun/8.12.8) with ESMTP id hAE9cX8x014951 for ; Fri, 14 Nov 2003 20:38:33 +1100 (EST) Received: from localhost (sgaric@localhost) by bruce.csse.monash.edu.au (8.12.8+Sun/8.12.2/Submit) with ESMTP id hAE9cXtu014948 for ; Fri, 14 Nov 2003 20:38:33 +1100 (EST) Date: Fri, 14 Nov 2003 20:38:33 +1100 (EST) From: Slavisa Garic Subject: INSERT extremely slow with large data sets (fwd) To: pgsql-performance@postgresql.org Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Content-transfer-encoding: 7BIT X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/106 X-Sequence-Number: 4666 Hi Everyone, I am using PostgreSQL 7.3.2 and have used earlier versions (7.1.x onwards) and with all of them I noticed same problem with INSERTs when there is a large data set. Just to so you guys can compare time it takes to insert one row into a table when there are only few rows present and when there are thousands: Rows Present Start Time Finish Time ------------------------------------------------------------ 100 1068790804.12 1068790804.12 1000 1068790807.87 1068790807.87 5000 1068790839.26 1068790839.27 10000 1068790909.24 1068790909.26 20000 1068791172.82 1068791172.85 30000 1068791664.06 1068791664.09 40000 1068792369.94 1068792370.0 50000 1068793317.53 1068793317.6 60000 1068794369.38 1068794369.47 As you can see if takes awfully lots of time for me just to have those values inserted. Now to make a picture a bit clearer for you this table has lots of information in there, about 25 columns. Also there are few indexes that I created so that the process of selecting values from there is faster which by the way works fine. Selecting anything takes under 5 seconds. Any help would be greatly appreciated even pointing me in the right direction where to ask this question. By the way I designed the database this way as my application that uses PGSQL a lot during the execution so there was a huge need for fast SELECTs. Our experiments are getting larger and larger every day so fast inserts would be good as well. Just to note those times above are of INSERTs only. Nothing else done that would be included in those times. Machine was also free and that was the only process running all the time and the machine was Intel(R) Pentium(R) 4 CPU 2.40GHz. Regards, Slavisa From pgsql-performance-owner@postgresql.org Fri Nov 14 06:12:40 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DABDFD1B544 for ; Fri, 14 Nov 2003 10:12:37 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 54100-10 for ; Fri, 14 Nov 2003 06:12:09 -0400 (AST) Received: from easily.co.uk (unknown [213.161.76.90]) by svr1.postgresql.org (Postfix) with ESMTP id 8617BD1B535 for ; Fri, 14 Nov 2003 06:12:04 -0400 (AST) Received: from [213.152.63.90] (HELO webbased10) by easily.co.uk (CommuniGate Pro SMTP 4.1.3) with SMTP id 34415623 for pgsql-performance@postgresql.org; Fri, 14 Nov 2003 10:11:46 +0000 Message-ID: <002201c3aa97$b53bbac0$2802a8c0@webbased10> From: "Nick Barr" To: Subject: IDE Hardware RAID Controller Date: Fri, 14 Nov 2003 10:11:44 -0000 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests= X-Spam-Level: X-Archive-Number: 200311/107 X-Sequence-Number: 4667 Heya, FYI just spotted this and thought I would pass it on, for all those who are looking at new boxes. http://www.theinquirer.net/?article=12665 http://www.promise.com/product/product_detail_eng.asp?productId=112&familyId =2 Looks like a four-channel hot-swap IDE (SATA) hardware RAID controller with up to 256Mb onboard RAM. Nick From pgsql-performance-owner@postgresql.org Fri Nov 14 09:38:45 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 92DE4D1B50D for ; Fri, 14 Nov 2003 13:38:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 80821-09 for ; Fri, 14 Nov 2003 09:38:12 -0400 (AST) Received: from jefftrout.com (h00a0cc4084e5.ne.client2.attbi.com [24.128.241.68]) by svr1.postgresql.org (Postfix) with SMTP id 0D2A3D1B51C for ; Fri, 14 Nov 2003 09:38:08 -0400 (AST) Received: (qmail 90440 invoked from network); 14 Nov 2003 13:38:10 -0000 Received: from localhost (HELO squeegit) (threshar@127.0.0.1) by localhost with SMTP; 14 Nov 2003 13:38:10 -0000 Date: Fri, 14 Nov 2003 08:37:55 -0500 From: Jeff To: Slavisa Garic Cc: pgsql-performance@postgresql.org Subject: Re: INSERT extremely slow with large data sets (fwd) Message-Id: <20031114083755.67841110.threshar@torgo.978.org> In-Reply-To: References: X-Mailer: Sylpheed version 0.9.7 (GTK+ 1.2.10; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/108 X-Sequence-Number: 4668 On Fri, 14 Nov 2003 20:38:33 +1100 (EST) Slavisa Garic wrote: > Any help would be greatly appreciated even pointing me in the right > direction where to ask this question. By the way I designed the > database this way as my application that uses PGSQL a lot during the > execution so there was a huge need for fast SELECTs. Our experiments > are getting larger and larger every day so fast inserts would be good > as well. > First, you need to upgrade to 7.3.4, 7.4 is prefable if a dump/reload is not too bad. Standard set of questions: 1. Any foreign keys 2. are these inserts batched into transactions 3. CPU usage? 4. OS? 5. PG config? [shared_buffers, effective_cache_size, etc] 6. IO saturation? Also, try searching the archives. lots of juicy info there too. -- Jeff Trout http://www.jefftrout.com/ http://www.stuarthamm.net/ From pgsql-performance-owner@postgresql.org Fri Nov 14 10:49:32 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3B953D1B51E for ; Fri, 14 Nov 2003 14:49:26 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 97934-05 for ; Fri, 14 Nov 2003 10:48:58 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id C557AD1B540 for ; Fri, 14 Nov 2003 10:48:54 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hAEEmsNu027454 for ; Fri, 14 Nov 2003 14:48:54 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hAEEKNmv021905 for pgsql-performance@postgresql.org; Fri, 14 Nov 2003 14:20:23 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Help with count(*) Date: Fri, 14 Nov 2003 09:13:48 -0500 Organization: cbbrowne Computing Inc Lines: 19 Message-ID: References: <200311141251.38786.mallah@trade-india.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:+5cNivcH5pqcGL+Z9tMQuNk0Vto= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/109 X-Sequence-Number: 4669 Martha Stewart called it a Good Thing when mallah@trade-india.com (Rajesh Kumar Mallah) wrote: > INFO: "profiles": found 0 removable, 369195 nonremovable row versions in 43423 pages > DETAIL: 246130 dead row versions cannot be removed yet. > Nonremovable row versions range from 136 to 2036 bytes long. It seems as though you have a transaction open that is holding onto a whole lot of old rows. I have seen this happen somewhat-invisibly when a JDBC connection manager opens transactions for each connection, and then no processing happens to use those connections for a long time. The open transactions prevent vacuums from doing any good... -- If this was helpful, rate me http://cbbrowne.com/info/multiplexor.html "Waving away a cloud of smoke, I look up, and am blinded by a bright, white light. It's God. No, not Richard Stallman, or Linus Torvalds, but God. In a booming voice, He says: "THIS IS A SIGN. USE LINUX, THE FREE Unix SYSTEM FOR THE 386." -- Matt Welsh From pgsql-performance-owner@postgresql.org Fri Nov 14 11:17:03 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CD503D1B55B for ; Fri, 14 Nov 2003 15:10:39 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 01129-03 for ; Fri, 14 Nov 2003 11:10:12 -0400 (AST) Received: from mystic.root.hu (mystic.root.hu [62.68.166.4]) by svr1.postgresql.org (Postfix) with ESMTP id 8203FD1B50B for ; Fri, 14 Nov 2003 11:10:07 -0400 (AST) Received: from [213.163.10.103] (helo=fejleszt4) by mystic.root.hu with asmtp (Exim 3.36 #1 (Debian)) id 1AKfaC-0007Qh-00 (authenticated sender: fac545198474dd6187837a07256e3b3f) for ; Fri, 14 Nov 2003 16:10:00 +0100 Message-ID: <01a601c3aac1$281e72c0$0403a8c0@fejleszt4> From: "=?iso-8859-2?B?U1rbQ1MgR+Fib3I=?=" To: Subject: constant vs function param differs in performance Date: Fri, 14 Nov 2003 16:08:27 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-2" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Authenticated-Sender: fac545198474dd6187837a07256e3b3f X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests= X-Spam-Level: X-Archive-Number: 200311/110 X-Sequence-Number: 4670 Dear Gurus, I have two SQL function that produce different times and I can't understand why. Here is the basic difference between them: CREATE FUNCTION test_const_1234 () RETURNS int4 AS ' SELECT ... 1234 ... 1234 .... 1234 ... ' LANGUAGE 'SQL'; CREATE FUNCTION test_param (int4) RETURNS int4 AS ' SELECT ... $1 .... $1 .... $1 ... ' LANGUAGE 'SQL'; Some sample times for different data: test_const_1234() 450 msec test_param(1234) 2700-4000 msec (probably disk cache) test_const_5678() 13500 msec test_param(5678) 14500 msec Is there a sane explanation? a solution? I can send more info if you wish. TIA, G. ------------------------------- cut here ------------------------------- From pgsql-performance-owner@postgresql.org Fri Nov 14 12:01:56 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5A3FBD1B4E4 for ; Fri, 14 Nov 2003 16:01:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 10469-04 for ; Fri, 14 Nov 2003 12:01:15 -0400 (AST) Received: from shire.ontko.com (shire.ontko.com [199.164.165.1]) by svr1.postgresql.org (Postfix) with ESMTP id 1AF4BD1B50C for ; Fri, 14 Nov 2003 12:01:15 -0400 (AST) Received: from nick (bilbo.ontko.com [199.164.165.101]) by shire.ontko.com (8.12.3/8.12.3/Debian-6.6) with SMTP id hAEG0n5f024840; Fri, 14 Nov 2003 11:00:50 -0500 Reply-To: From: "Nick Fankhauser" To: "Manfred Koizar" Cc: "Tom Lane" , "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long Date: Fri, 14 Nov 2003 11:00:38 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal In-Reply-To: X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/111 X-Sequence-Number: 4671 > Does actor_case_assignment contain more columns than just the two ids? > If yes, do these additional fields account for ca. 70 bytes per tuple? > If not, try > VACUUM FULL ANALYSE actor_case_assignment; actor_case_assignment has its own primary key and a "role" field in addition to the ids you've seen, so 70 bytes sounds reasonable. (The PK is to allow a remote mirroring application to update these records- otherwise it would be unnecessary.) > 7ms per > tuple returned looks like a lot of disk seeks are involved. Is > clustering actor on actor_full_name_uppercase an option or would this > slow down other queries? Good question... I've never used clustering in PostgreSQL before, so I'm unsure. I presume this is like clustering in Oracle where the table is ordered to match the index? If so, I think you may be onto something because the only other field We regularly query on is the actor_id. Actor_id has a unique index with no clustering currently, so I don't think I'd lose a thing by clustering on actor_full_name_uppercase. I'll give this a try & let you know how it changes. BTW, you are correct that caching has a big affect on the actual time figures in this case- I'm working on my development DB, so cahced info doesn't get trampled as quickly by other users. Is there a way to flush out the cache in a testing situation like this in order to start from a consistent base? Thanks! -Nick From pgsql-performance-owner@postgresql.org Thu Nov 13 12:34:59 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 7D63CD1B51C for ; Thu, 13 Nov 2003 16:01:59 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 67143-01 for ; Thu, 13 Nov 2003 12:01:28 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id 59328D1B552 for ; Thu, 13 Nov 2003 12:01:27 -0400 (AST) Received: (qmail 2473 invoked from network); 13 Nov 2003 16:02:34 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 13 Nov 2003 16:02:34 -0000 Message-ID: <3FB4FC66.3060503@trade-india.com> Date: Fri, 14 Nov 2003 21:31:42 +0530 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Robert Treat Cc: pgsql-performance@postgresql.org Subject: Re: IN surpasses NOT EXISTS in 7.4RC2 ?? References: <200311131323.39762.mallah@trade-india.com> <1068737502.10946.18338.camel@camel> In-Reply-To: <1068737502.10946.18338.camel@camel> Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/89 X-Sequence-Number: 4649 Robert Treat wrote:
It is believed that the IN optimization can lead to faster IN times than
EXIST times on some queries, the extent of which is still a bit of an
unknown. (Incidentally is there an FAQ item on this that needs
updating?)

Thanks every one for clarifying. Its really a nice thing to see IN working
so well becoz its easier to read the SQL using IN.

looks like NOT IN is indifferent to indexes where is IN uses indexes , is it true?

does indexes affect the new manner in which IN works in 7.4 ?





Does the not exist query produce worse results in 7.4 than it did in
7.3?
Will surely post the overvation sometime.



Regards
Mallah.



Robert Treat

On Thu, 2003-11-13 at 02:53, Rajesh Kumar Mallah wrote:
  
Hi,

NOT EXISTS is taking almost double time than NOT IN .
I know IN has been optimised in 7.4 but is anything 
wrong with the NOT EXISTS?

I have vaccumed , analyze and run the query many times
still not in is faster than exists :>


Regds
Mallah.

NOT IN PLAN

tradein_clients=# explain analyze SELECT  count(*) from general.profile_master where
 profile_id not in (select  profile_id from general.account_profiles ) ;
                                                             QUERY PLAN
-------------------------------------------------------------------------------------------------------------------------------------
Aggregate  (cost=32238.19..32238.19 rows=1 width=0) (actual time=5329.206..5329.207 rows=1 loops=1)
  ->  Seq Scan on profile_master  (cost=4458.25..31340.38 rows=359125 width=0) (actual time=1055.496..4637.908 rows=470386 loops=1)
        Filter: (NOT (hashed subplan))
        SubPlan
          ->  Seq Scan on account_profiles  (cost=0.00..3817.80 rows=256180 width=4) (actual time=0.061..507.811 rows=256180 loops=1)
Total runtime: 5337.591 ms
(6 rows)


tradein_clients=# explain analyze SELECT  count(*) from general.profile_master where not exists 
(select  profile_id from general.account_profiles where profile_id=general.profile_master.profile_id ) ;

                                                                          QUERY PLAN
---------------------------------------------------------------------------------------------------------------------------------------------------------------
Aggregate  (cost=1674981.97..1674981.97 rows=1 width=0) (actual time=14600.386..14600.387 rows=1 loops=1)
  ->  Seq Scan on profile_master  (cost=0.00..1674084.16 rows=359125 width=0) (actual time=13.687..13815.798 rows=470386 loops=1)
        Filter: (NOT (subplan))
        SubPlan
          ->  Index Scan using account_profiles_profile_id on account_profiles  (cost=0.00..4.59 rows=2 width=4) (actual time=0.013..0.013 rows=0 loops=718250)
                Index Cond: (profile_id = $0)
Total runtime: 14600.531 ms



---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
    (send "unregister YourEmailAddressHere" to majordomo@postgresql.org)

From pgsql-performance-owner@postgresql.org Fri Nov 14 12:08:18 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D59A6D1B55C for ; Fri, 14 Nov 2003 16:08:10 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 12108-02 for ; Fri, 14 Nov 2003 12:07:39 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id 2A50AD1B50D for ; Fri, 14 Nov 2003 12:07:36 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 9DF361FAE; Fri, 14 Nov 2003 11:07:15 -0500 (EST) To: Josh Berkus Cc: Christopher Kings-Lynne , PostgreSQL Performance Subject: Re: Query question From: Neil Conway In-Reply-To: <200311122120.28882.josh@agliodbs.com> (Josh Berkus's message of "Wed, 12 Nov 2003 21:20:28 -0800") References: <3FB303E9.20903@familyhealth.com.au> <200311122120.28882.josh@agliodbs.com> Date: Fri, 14 Nov 2003 11:07:15 -0500 Message-ID: <87llqjj5n0.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/112 X-Sequence-Number: 4672 Josh Berkus writes: > The only thing you're adding to the query is a second SORT step, so it > shouldn't require any more time/memory than the query's first SORT > did. Interesting -- I wonder if it would be possible for the optimizer to detect this and avoid the redundant inner sort ... (/me muses to himself) -Neil From pgsql-performance-owner@postgresql.org Fri Nov 14 12:38:26 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A5731D1B535 for ; Fri, 14 Nov 2003 16:38:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 18783-06 for ; Fri, 14 Nov 2003 12:37:53 -0400 (AST) Received: from email04.aon.at (WARSL402PIP2.highway.telekom.at [195.3.96.74]) by svr1.postgresql.org (Postfix) with SMTP id CDD2AD1B50D for ; Fri, 14 Nov 2003 12:37:50 -0400 (AST) Received: (qmail 313860 invoked from network); 14 Nov 2003 16:37:43 -0000 Received: from m170p014.dipool.highway.telekom.at (HELO cantor) ([62.46.11.46]) (envelope-sender ) by qmail7rs.highway.telekom.at (qmail-ldap-1.03) with SMTP for ; 14 Nov 2003 16:37:43 -0000 From: Manfred Koizar To: Cc: "Tom Lane" , "Pgsql-Performance@Postgresql. Org" Subject: Re: Seeking help with a query that takes too long Date: Fri, 14 Nov 2003 17:35:44 +0100 Message-ID: References: In-Reply-To: X-Mailer: Forte Agent 1.93/32.576 English (American) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/113 X-Sequence-Number: 4673 On Fri, 14 Nov 2003 11:00:38 -0500, "Nick Fankhauser" wrote: >Good question... I've never used clustering in PostgreSQL before, so I'm >unsure. I presume this is like clustering in Oracle where the table is >ordered to match the index? Yes, something like that. With the exception that Postgres looses the clustered status, while you INSERT and UPDATE tuples. So you have to re-CLUSTER from time to time. Look at pg_stats.correlation to see, if its necessary. > Is there a way to flush out >the cache in a testing situation like this in order to start from a >consistent base? To flush Postgres shared buffers: SELECT count(*) FROM another_large_table; To flush your database pages from the OS cache: tar cf /dev/null /some/large/directory And run each of your tests at least twice to get a feeling how caching affects your specific queries. Servus Manfred From pgsql-performance-owner@postgresql.org Fri Nov 14 12:41:28 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 850F1D1B50C for ; Fri, 14 Nov 2003 16:41:26 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 18783-07 for ; Fri, 14 Nov 2003 12:40:55 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 58C84D1B535 for ; Fri, 14 Nov 2003 12:40:54 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAEGeh19026649; Fri, 14 Nov 2003 11:40:43 -0500 (EST) To: Neil Conway Cc: Josh Berkus , Christopher Kings-Lynne , PostgreSQL Performance Subject: Re: Query question In-reply-to: <87llqjj5n0.fsf@mailbox.samurai.com> References: <3FB303E9.20903@familyhealth.com.au> <200311122120.28882.josh@agliodbs.com> <87llqjj5n0.fsf@mailbox.samurai.com> Comments: In-reply-to Neil Conway message dated "Fri, 14 Nov 2003 11:07:15 -0500" Date: Fri, 14 Nov 2003 11:40:42 -0500 Message-ID: <26648.1068828042@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/114 X-Sequence-Number: 4674 Neil Conway writes: > Interesting -- I wonder if it would be possible for the optimizer to > detect this and avoid the redundant inner sort ... (/me muses to > himself) I think the ability to generate two sort steps is a feature, not a bug. This has been often requested in connection with user-defined aggregates, where it's handy to be able to control the order of arrival of rows at the aggregation function. If the optimizer suppressed the inner sort then we'd lose that ability. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Nov 14 13:44:00 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D6801D1B53F for ; Fri, 14 Nov 2003 17:43:57 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 29579-04 for ; Fri, 14 Nov 2003 13:43:27 -0400 (AST) Received: from fuji.krosing.net (silmet.estpak.ee [194.126.97.78]) by svr1.postgresql.org (Postfix) with ESMTP id 4D01DD1B51C for ; Fri, 14 Nov 2003 13:43:26 -0400 (AST) Received: from fuji.krosing.net (localhost.localdomain [127.0.0.1]) by fuji.krosing.net (8.12.8/8.12.8) with ESMTP id hAEHhSP5003241; Fri, 14 Nov 2003 19:43:28 +0200 Received: (from hannu@localhost) by fuji.krosing.net (8.12.8/8.12.8/Submit) id hAEHhSZd003239; Fri, 14 Nov 2003 19:43:28 +0200 X-Authentication-Warning: fuji.krosing.net: hannu set sender to hannu@tm.ee using -f Subject: Re: Help with count(*) From: Hannu Krosing To: Christopher Browne Cc: pgsql-performance@postgresql.org In-Reply-To: References: <200311141251.38786.mallah@trade-india.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1068831807.2540.44.camel@fuji.krosing.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Fri, 14 Nov 2003 19:43:27 +0200 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/115 X-Sequence-Number: 4675 Christopher Browne kirjutas R, 14.11.2003 kell 16:13: > Martha Stewart called it a Good Thing when mallah@trade-india.com (Rajesh Kumar Mallah) wrote: > > INFO: "profiles": found 0 removable, 369195 nonremovable row versions in 43423 pages > > DETAIL: 246130 dead row versions cannot be removed yet. > > Nonremovable row versions range from 136 to 2036 bytes long. > > It seems as though you have a transaction open that is holding onto a > whole lot of old rows. > > I have seen this happen somewhat-invisibly when a JDBC connection > manager opens transactions for each connection, and then no processing > happens to use those connections for a long time. The open > transactions prevent vacuums from doing any good... Can't the backend be made to delay the "real" start of transaction until the first query gets executed ? ------------ Hannu From pgsql-performance-owner@postgresql.org Fri Nov 14 14:51:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9A0F3D1B518 for ; Fri, 14 Nov 2003 18:51:45 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 42993-08 for ; Fri, 14 Nov 2003 14:51:15 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id A991AD1B502 for ; Fri, 14 Nov 2003 14:51:13 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAEInx19003132; Fri, 14 Nov 2003 13:49:59 -0500 (EST) To: Hannu Krosing Cc: Christopher Browne , pgsql-performance@postgresql.org Subject: Re: Help with count(*) In-reply-to: <1068831807.2540.44.camel@fuji.krosing.net> References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> Comments: In-reply-to Hannu Krosing message dated "Fri, 14 Nov 2003 19:43:27 +0200" Date: Fri, 14 Nov 2003 13:49:59 -0500 Message-ID: <3131.1068835799@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/116 X-Sequence-Number: 4676 Hannu Krosing writes: > Christopher Browne kirjutas R, 14.11.2003 kell 16:13: >> I have seen this happen somewhat-invisibly when a JDBC connection >> manager opens transactions for each connection, and then no processing >> happens to use those connections for a long time. The open >> transactions prevent vacuums from doing any good... > Can't the backend be made to delay the "real" start of transaction until > the first query gets executed ? That is on the TODO list. I looked at it briefly towards the end of the 7.4 development cycle, and decided that it was nontrivial and I didn't have time to make it happen before beta started. I don't recall why it didn't seem trivial. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Nov 14 15:49:33 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0649ED1B4E1 for ; Fri, 14 Nov 2003 19:49:31 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 53927-01 for ; Fri, 14 Nov 2003 15:49:01 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 47F45D1B4E4 for ; Fri, 14 Nov 2003 15:49:00 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hAEJn0Nu046147 for ; Fri, 14 Nov 2003 19:49:00 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hAEJKfCQ041683 for pgsql-performance@postgresql.org; Fri, 14 Nov 2003 19:20:41 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Help with count(*) Date: Fri, 14 Nov 2003 14:16:56 -0500 Organization: cbbrowne Computing Inc Lines: 35 Message-ID: References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:vI2O3yzZC8Gyhd5OMVqxPwSIIdc= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/118 X-Sequence-Number: 4678 After a long battle with technology, hannu@tm.ee (Hannu Krosing), an earthling, wrote: > Christopher Browne kirjutas R, 14.11.2003 kell 16:13: >> Martha Stewart called it a Good Thing when mallah@trade-india.com (Rajesh Kumar Mallah) wrote: >> > INFO: "profiles": found 0 removable, 369195 nonremovable row versions in 43423 pages >> > DETAIL: 246130 dead row versions cannot be removed yet. >> > Nonremovable row versions range from 136 to 2036 bytes long. >> >> It seems as though you have a transaction open that is holding onto a >> whole lot of old rows. >> >> I have seen this happen somewhat-invisibly when a JDBC connection >> manager opens transactions for each connection, and then no processing >> happens to use those connections for a long time. The open >> transactions prevent vacuums from doing any good... > > Can't the backend be made to delay the "real" start of transaction until > the first query gets executed ? One would hope so. Some time when I have the Round Tuits, I ought to take a browse of the connection pool code to notice if there's anything to notice. The thing that I keep imagining would be a slick idea would be to have a thread periodically go through once for however many connections the pool permits and fire a short transaction through every otherwise-unoccupied connection in the pool, in effect, doing a sort of "vacuum" of the connections. I don't get very favorable reactions when I suggest that, though... -- (reverse (concatenate 'string "ac.notelrac.teneerf" "@" "454aa")) http://cbbrowne.com/info/sgml.html Rules of the Evil Overlord #80. "If my weakest troops fail to eliminate a hero, I will send out my best troops instead of wasting time with progressively stronger ones as he gets closer and closer to my fortress." From pgsql-performance-owner@postgresql.org Thu Nov 13 15:27:58 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 25493D1B4E1 for ; Thu, 13 Nov 2003 19:27:58 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 16910-01 for ; Thu, 13 Nov 2003 15:27:28 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id 8A329D1B551 for ; Thu, 13 Nov 2003 15:27:26 -0400 (AST) Received: (qmail 10415 invoked from network); 13 Nov 2003 19:28:33 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 13 Nov 2003 19:28:33 -0000 Message-ID: <3FB52CAD.9070404@trade-india.com> Date: Sat, 15 Nov 2003 00:57:41 +0530 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Tom Lane Cc: pgsql-performance@postgresql.org Subject: Re: IN surpasses NOT EXISTS in 7.4RC2 ?? References: <200311131323.39762.mallah@trade-india.com> <26331.1068736791@sss.pgh.pa.us> In-Reply-To: <26331.1068736791@sss.pgh.pa.us> Content-Type: multipart/alternative; boundary="------------070600040606020801020106" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/98 X-Sequence-Number: 4658 This is a multi-part message in MIME format. --------------070600040606020801020106 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Tom Lane wrote: >Rajesh Kumar Mallah writes: > > >>NOT EXISTS is taking almost double time than NOT IN . >>I know IN has been optimised in 7.4 but is anything >>wrong with the NOT EXISTS? >> >> > >That's the expected behavior in 7.4. EXISTS in the style you are using >it effectively forces a nestloop-with-inner-indexscan implementation. >As of 7.4, IN can do that, but it can do several other things too, >including the hash-type plan you have here. So assuming that the >planner chooses the right plan choice (not always a given ;-)) > >IN should be as fast or faster than EXISTS in *all* *cases.* > Not in this case :) , did i miss something silly? tradein_clients=# explain SELECT count(*) from user_accounts where email is not null and email not in (select email from profile_master where email is not null) ; QUERY PLAN -------------------------------------------------------------------------------------- Aggregate (cost=*9587726326.93..9587726326.93* rows=1 width=0) -> Seq Scan on user_accounts (cost=0.00..9587725473.40 rows=341412 width=0) Filter: ((email IS NOT NULL) AND (NOT (subplan))) SubPlan -> Seq Scan on profile_master (cost=0.00..25132.24 rows=674633 width=25) Filter: (email IS NOT NULL) (6 rows) *The query above does not return* tradein_clients=# explain analyze SELECT count(*) from user_accounts where email is not null and not exists (select email from profile_master where email=user_accounts.email) ; QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------------------------------------- Aggregate (cost=2850847.55..2850847.55 rows=1 width=0) (actual time=34075.100..34075.101 rows=1 loops=1) -> Seq Scan on user_accounts (cost=0.00..2849994.02 rows=341412 width=0) (actual time=8.066..34066.329 rows=3882 loops=1) Filter: ((email IS NOT NULL) AND (NOT (subplan))) SubPlan -> Index Scan using profile_master_email on profile_master (cost=0.00..35.60 rows=9 width=25) (actual time=0.044..0.044 rows=1 loops=686716) Index Cond: ((email)::text = ($0)::text) Total runtime: 34075.213 ms (7 rows) tradein_clients=# > > regards, tom lane > >---------------------------(end of broadcast)--------------------------- >TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org > > -- Rajesh Kumar Mallah, Infocom Network Limited, New Delhi phone: +91(11)6152172 (221) (L) ,9811255597 (M) Visit http://www.trade-india.com , India's Leading B2B eMarketplace. --------------070600040606020801020106 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Tom Lane wrote:
Rajesh Kumar Mallah <mallah@trade-india.com> writes:
  
NOT EXISTS is taking almost double time than NOT IN .
I know IN has been optimised in 7.4 but is anything 
wrong with the NOT EXISTS?
    

That's the expected behavior in 7.4.  EXISTS in the style you are using
it effectively forces a nestloop-with-inner-indexscan implementation.
As of 7.4, IN can do that, but it can do several other things too,
including the hash-type plan you have here.  So assuming that the
planner chooses the right plan choice (not always a given ;-))

IN should be as fast or faster than EXISTS in all cases.

Not in this case :) , did i miss something silly?

tradein_clients=# explain  SELECT count(*) from user_accounts where email is not null and email not in
 (select email from profile_master where email is not null) ;
                                      QUERY PLAN
--------------------------------------------------------------------------------------
 Aggregate  (cost=9587726326.93..9587726326.93 rows=1 width=0)
   ->  Seq Scan on user_accounts  (cost=0.00..9587725473.40 rows=341412 width=0)
         Filter: ((email IS NOT NULL) AND (NOT (subplan)))
         SubPlan
           ->  Seq Scan on profile_master  (cost=0.00..25132.24 rows=674633 width=25)
                 Filter: (email IS NOT NULL)
(6 rows)

The query above does not return

tradein_clients=# explain analyze SELECT count(*) from user_accounts where email is not null and
not exists (select email from profile_master where email=user_accounts.email) ;
                                                                        QUERY PLAN
----------------------------------------------------------------------------------------------------------------------------------------------------------
 Aggregate  (cost=2850847.55..2850847.55 rows=1 width=0) (actual time=34075.100..34075.101 rows=1 loops=1)
   ->  Seq Scan on user_accounts  (cost=0.00..2849994.02 rows=341412 width=0) (actual time=8.066..34066.329 rows=3882 loops=1)
         Filter: ((email IS NOT NULL) AND (NOT (subplan)))
         SubPlan
           ->  Index Scan using profile_master_email on profile_master  (cost=0.00..35.60 rows=9 width=25) (actual time=0.044..0.044 rows=1 loops=686716)
                 Index Cond: ((email)::text = ($0)::text)
 Total runtime: 34075.213 ms
(7 rows)

tradein_clients=#



			regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
  


-- 

Rajesh Kumar Mallah,
Infocom Network Limited, New Delhi
phone: +91(11)6152172 (221) (L) ,9811255597 (M)

Visit http://www.trade-india.com ,
India's Leading B2B eMarketplace.
--------------070600040606020801020106-- From pgsql-performance-owner@postgresql.org Fri Nov 14 15:41:02 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 26C03D1B4E4 for ; Fri, 14 Nov 2003 19:41:01 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 51358-10 for ; Fri, 14 Nov 2003 15:40:30 -0400 (AST) Received: from jwilhelm.ofsloans.com (unknown [209.180.142.225]) by svr1.postgresql.org (Postfix) with ESMTP id 78FE8D1B50B for ; Fri, 14 Nov 2003 15:40:28 -0400 (AST) Message-ID: <3FB52FA7.8060508@lashell.net> Date: Fri, 14 Nov 2003 12:40:23 -0700 From: Will LaShell X-Accept-Language: en-us, en MIME-Version: 1.0 To: Hannu Krosing Cc: Christopher Browne , pgsql-performance@postgresql.org Subject: Re: Help with count(*) References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> In-Reply-To: <1068831807.2540.44.camel@fuji.krosing.net> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/117 X-Sequence-Number: 4677 Hannu Krosing wrote: >Christopher Browne kirjutas R, 14.11.2003 kell 16:13: > > >>Martha Stewart called it a Good Thing when mallah@trade-india.com (Rajesh Kumar Mallah) wrote: >> >> >>>INFO: "profiles": found 0 removable, 369195 nonremovable row versions in 43423 pages >>>DETAIL: 246130 dead row versions cannot be removed yet. >>>Nonremovable row versions range from 136 to 2036 bytes long. >>> >>> >>It seems as though you have a transaction open that is holding onto a >>whole lot of old rows. >> >>I have seen this happen somewhat-invisibly when a JDBC connection >>manager opens transactions for each connection, and then no processing >>happens to use those connections for a long time. The open >>transactions prevent vacuums from doing any good... >> >> > >Can't the backend be made to delay the "real" start of transaction until >the first query gets executed ? > > That seems counter intuitive doesn't it? Why write more code in the server when the client is the thing that has the problem? Will From pgsql-performance-owner@postgresql.org Fri Nov 14 16:39:27 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1149CD1B4EB for ; Fri, 14 Nov 2003 20:39:23 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 62492-03 for ; Fri, 14 Nov 2003 16:38:53 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id B374AD1B4E1 for ; Fri, 14 Nov 2003 16:38:47 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAEKck19004988; Fri, 14 Nov 2003 15:38:47 -0500 (EST) To: Will LaShell Cc: Hannu Krosing , Christopher Browne , pgsql-performance@postgresql.org Subject: Re: Help with count(*) In-reply-to: <3FB52FA7.8060508@lashell.net> References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> Comments: In-reply-to Will LaShell message dated "Fri, 14 Nov 2003 12:40:23 -0700" Date: Fri, 14 Nov 2003 15:38:46 -0500 Message-ID: <4987.1068842326@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/119 X-Sequence-Number: 4679 Will LaShell writes: > Hannu Krosing wrote: >> Can't the backend be made to delay the "real" start of transaction until >> the first query gets executed ? > That seems counter intuitive doesn't it? Why write more code in the > server when the client is the thing that has the problem? Because there are a lot of clients with the same problem :-( A more principled argument is that we already postpone the setting of the transaction snapshot until the first query arrives within the transaction. In a very real sense, the setting of the snapshot *is* the start of the transaction. So it would make sense if incidental stuff like VACUUM also thought that the transaction hadn't started until the first query arrives. (I believe the previous discussion also agreed that we wanted to postpone the freezing of now(), which currently also happens at BEGIN rather than the first command after BEGIN.) regards, tom lane From pgsql-performance-owner@postgresql.org Fri Nov 14 16:59:47 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A4278D1B4ED for ; Fri, 14 Nov 2003 20:59:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 55314-10 for ; Fri, 14 Nov 2003 16:59:16 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id A1D5FD1B4EB for ; Fri, 14 Nov 2003 16:59:14 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAEKxE19005188; Fri, 14 Nov 2003 15:59:14 -0500 (EST) To: "=?iso-8859-2?B?U1rbQ1MgR+Fib3I=?=" Cc: pgsql-performance@postgresql.org Subject: Re: constant vs function param differs in performance In-reply-to: <01a601c3aac1$281e72c0$0403a8c0@fejleszt4> References: <01a601c3aac1$281e72c0$0403a8c0@fejleszt4> Comments: In-reply-to "=?iso-8859-2?B?U1rbQ1MgR+Fib3I=?=" message dated "Fri, 14 Nov 2003 16:08:27 +0100" Date: Fri, 14 Nov 2003 15:59:14 -0500 Message-ID: <5187.1068843554@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/120 X-Sequence-Number: 4680 "=?iso-8859-2?B?U1rbQ1MgR+Fib3I=?=" writes: > I have two SQL function that produce different times and I can't understand > why. The planner often produces different plans when there are constants in WHERE clauses than when there are variables, because it can get more accurate ideas of how many rows will be retrieved. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Nov 14 17:28:00 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3781BD1B4ED for ; Fri, 14 Nov 2003 21:27:59 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 72153-01 for ; Fri, 14 Nov 2003 17:27:30 -0400 (AST) Received: from shire.ontko.com (shire.ontko.com [199.164.165.1]) by svr1.postgresql.org (Postfix) with ESMTP id 0DCD8D1B51A for ; Fri, 14 Nov 2003 17:27:28 -0400 (AST) Received: from nick (bilbo.ontko.com [199.164.165.101]) by shire.ontko.com (8.12.3/8.12.3/Debian-6.6) with SMTP id hAELRR5f030427; Fri, 14 Nov 2003 16:27:29 -0500 Reply-To: From: "Nick Fankhauser" To: "Pgsql-Performance@Postgresql. Org" Cc: "Ray Ontko" Subject: n_distinct way off, but following a pattern. Date: Fri, 14 Nov 2003 16:27:16 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/121 X-Sequence-Number: 4681 Hi- I'm seeing estimates for n_distinct that are way off for a large table (8,700,000 rows). They get better by setting the stats target higher, but are still off by a factor of 10 with the stats set to 1000. I've noticed and reported a similar pattern before on another table. Because this follows the same very consistent pattern, I figured it was worth reporting again. This looks more like a bug than randomness. If the poor result was simply due to having a small sample to work from, the estimates should be all over the map, but these are consistently low, and vary in almost exact inverse proportion to the stats target: run 1: run2: run3: n_distinct estimate, statistics = 10: 3168 3187 3212 n_distinct estimate, statistics = 100: 23828 24059 23615 n_distinct estimate, statistics = 1000: 194690 194516 194081 Actual distinct values: 3340724 Or to put it another way, if you were to take the estimate from analyze, divide by the stats target and multiply by 10000, the result would be pretty close to exact. (Within a factor of 2, which ought to be plenty close for planning purposes.) I'm running version 7.3.2 Any thoughts from folks familiar with this part of the source code? Regards, -Nick PS- Here's a log of the session that I got this from. alpha=# select count(distinct actor_id) from actor_case_assignment; -[ RECORD 1 ]-- count | 3340724 alpha=# analyze; ANALYZE alpha=# SELECT * FROM pg_stats alpha-# WHERE tablename='actor_case_assignment' AND attname='actor_id'; -[ RECORD 1 ]-----+------------------------------------------------------------------- ---------------------------------------------------------------------------- ---------------- schemaname | public tablename | actor_case_assignment attname | actor_id null_frac | 0 avg_width | 16 n_distinct | 3168 most_common_vals | {18105XS,18115XS,18106XS,18113JD02,18115JD02,18106J27,18113XS,18113A10656,18 115LST,18108XS} most_common_freqs | {0.0206667,0.0206667,0.0196667,0.019,0.0176667,0.0173333,0.0163333,0.015,0.0 14,0.0136667} histogram_bounds | {18067A000-07P,18067PD397SC1574-1,18105LBPD,18106A2119-49,18106PD399IF845-1, 18108A03068-20,18108LECS207,18108PTW03737278-2,18111A19788-77,18115A50420,18 115XC} correlation | 0.876795 alpha=# alpha=# alter table actor_case_assignment alter column actor_id set statistics 100; ALTER TABLE alpha=# analyze actor_case_assignment; ANALYZE alpha=# SELECT * FROM pg_stats alpha-# WHERE tablename='actor_case_assignment' AND attname='actor_id'; -[ RECORD 1 ]
schemaname | public tablename | actor_case_assignment attname | actor_id null_frac | 0 avg_width | 17 n_distinct | 23828 most_common_vals | {18115XS,18113JD02,18106XS,1.... alpha=# alter table actor_case_assignment alter column actor_id set statistics 1000; ALTER TABLE alpha=# analyze actor_case_assignment; ANALYZE alpha=# SELECT * FROM pg_stats alpha-# WHERE tablename='actor_case_assignment' AND attname='actor_id'; -[ RECORD 1 ]-----
schemaname | public tablename | actor_case_assignment attname | actor_id null_frac | 0 avg_width | 16 n_distinct | 194690 most_common_vals | {18106XS,18115XS,18115... alpha=# \x Expanded display is off. alpha=# alter table actor_case_assignment alter column actor_id set statistics 10; ALTER TABLE alpha=# analyze actor_case_assignment; ANALYZE alpha=# select n_distinct from pg_stats where tablename='actor_case_assignment and attname='actor_id'; alpha'# '; ERROR: parser: parse error at or near "actor_id" at character 85 alpha=# select n_distinct from pg_stats where tablename='actor_case_assignment' and attname='actor_id'; n_distinct ------------ 3187 (1 row) alpha=# alter table actor_case_assignment alter column actor_id set statistics 10; ALTER TABLE alpha=# analyze actor_case_assignment; ANALYZE alpha=# select n_distinct from pg_stats where tablename='actor_case_assignment' and attname='actor_id'; n_distinct ------------ 3212 (1 row) alpha=# analyze actor_case_assignment; ANALYZE alpha=# alter table actor_case_assignment alter column actor_id set statistics 100; ALTER TABLE alpha=# analyze actor_case_assignment; ANALYZE alpha=# select n_distinct from pg_stats where tablename='actor_case_assignment' and attname='actor_id'; n_distinct ------------ 24059 (1 row) alpha=# analyze actor_case_assignment; ANALYZE alpha=# alter table actor_case_assignment alter column actor_id set statistics 100; ALTER TABLE alpha=# select n_distinct from pg_stats where tablename='actor_case_assignment' and attname='actor_id'; n_distinct ------------ 23615 (1 row) alpha=# alter table actor_case_assignment alter column actor_id set statistics 1000; ALTER TABLE alpha=# analyze actor_case_assignment; ANALYZE alpha=# select n_distinct from pg_stats where tablename='actor_case_assignment' and attname='actor_id'; n_distinct ------------ 194516 (1 row) alpha=# analyze actor_case_assignment; ANALYZE alpha=# select n_distinct from pg_stats where tablename='actor_case_assignment' and attname='actor_id'; n_distinct ------------ 194081 (1 row) alpha=# From pgsql-performance-owner@postgresql.org Fri Nov 14 17:30:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 75140D1B563 for ; Fri, 14 Nov 2003 21:30:45 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 64706-06 for ; Fri, 14 Nov 2003 17:30:16 -0400 (AST) Received: from mail.libertyrms.com (unknown [209.167.124.227]) by svr1.postgresql.org (Postfix) with ESMTP id 3E473D1B570 for ; Fri, 14 Nov 2003 17:30:14 -0400 (AST) Received: from [10.1.2.130] (helo=dba2) by mail.libertyrms.com with esmtp (Exim 3.22 #3 (Debian)) id 1AKlWC-000711-00 for ; Fri, 14 Nov 2003 16:30:16 -0500 Received: by dba2 (Postfix, from userid 1019) id 0731BCBB8; Fri, 14 Nov 2003 16:30:16 -0500 (EST) Date: Fri, 14 Nov 2003 16:30:16 -0500 From: Andrew Sullivan To: pgsql-performance@postgresql.org Subject: Re: Help with count(*) Message-ID: <20031114213015.GT28153@libertyrms.info> Mail-Followup-To: Andrew Sullivan , pgsql-performance@postgresql.org References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/122 X-Sequence-Number: 4682 On Fri, Nov 14, 2003 at 02:16:56PM -0500, Christopher Browne wrote: > otherwise-unoccupied connection in the pool, in effect, doing a sort > of "vacuum" of the connections. I don't get very favorable reactions > when I suggest that, though... Because it's a kludge on top of another kludge, perhaps? ;-) This needs to be fixed properly, not through an ungraceful series of workarounds. A -- ---- Andrew Sullivan 204-4141 Yonge Street Afilias Canada Toronto, Ontario Canada M2P 2A8 +1 416 646 3304 x110 From pgsql-performance-owner@postgresql.org Fri Nov 14 18:03:09 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6EFBCD1B4F2 for ; Fri, 14 Nov 2003 22:03:08 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 73031-05 for ; Fri, 14 Nov 2003 18:02:39 -0400 (AST) Received: from web80210.mail.yahoo.com (web80210.mail.yahoo.com [66.218.79.45]) by svr1.postgresql.org (Postfix) with SMTP id 737EBD1B4E4 for ; Fri, 14 Nov 2003 18:02:36 -0400 (AST) Message-ID: <20031114220238.87764.qmail@web80210.mail.yahoo.com> Received: from [199.169.240.132] by web80210.mail.yahoo.com via HTTP; Fri, 14 Nov 2003 14:02:38 PST Date: Fri, 14 Nov 2003 14:02:38 -0800 (PST) From: George Essig Subject: Re: INSERT extremely slow with large data sets (fwd) To: Slavisa Garic , pgsql-performance@postgresql.org MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/123 X-Sequence-Number: 4683 Slavisa Garic wrote: > Hi Everyone, > I am using PostgreSQL 7.3.2 and have used earlier versions (7.1.x > onwards) > and with all of them I noticed same problem with INSERTs when there is > a > large data set. Just to so you guys can compare time it takes to insert > one row into a table when there are only few rows present and when > there > are thousands: Try running VACUUM ANALYZE periodically during inserts. I found this to help. George Essig From pgsql-novice-owner@postgresql.org Fri Nov 14 18:08:01 2003 X-Original-To: pgsql-novice-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1CCC0D1B53F; Fri, 14 Nov 2003 22:07:57 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 72910-07; Fri, 14 Nov 2003 18:07:28 -0400 (AST) Received: from smtp1.nodak.edu (smtp1.NoDak.edu [134.129.111.50]) by svr1.postgresql.org (Postfix) with ESMTP id 04109D1B4E4; Fri, 14 Nov 2003 18:07:25 -0400 (AST) Received: from ndsu.nodak.edu (webmail1.ndsu.NoDak.edu [134.129.111.141]) by smtp1.nodak.edu (8.11.6/8.11.6) with SMTP id hAEM7QB00739; Fri, 14 Nov 2003 16:07:26 -0600 Received: from 134.129.92.208 (SquirrelMail authenticated user radha.manohar) by webmail.ndsu.nodak.edu with HTTP; Fri, 14 Nov 2003 16:07:26 -0600 (CST) Message-ID: <1710.134.129.92.208.1068847646.squirrel@webmail.ndsu.nodak.edu> Date: Fri, 14 Nov 2003 16:07:26 -0600 (CST) Subject: Re: Error in transaction processing From: To: , X-Priority: 3 Importance: Normal X-Mailer: SquirrelMail (version 1.2.11) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/107 X-Sequence-Number: 8810 When I execute a transaction using embedded sql statements in a c program, I get the error, Error in transaction processing. I could see from the documentation that it means, "Postgres signalled to us that we cannot start, commit or rollback the transaction" I don't find any mistakes in the transaction statements. What can I do to correct this error? Your response would be very much appreciated. Thanks and Regards, Radha From pgsql-performance-owner@postgresql.org Fri Nov 14 18:13:06 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 35FBFD1B4E4 for ; Fri, 14 Nov 2003 22:13:05 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 81185-01 for ; Fri, 14 Nov 2003 18:12:36 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 7233FD1B561 for ; Fri, 14 Nov 2003 18:12:33 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAEMCS19005810; Fri, 14 Nov 2003 17:12:28 -0500 (EST) To: nickf@ontko.com Cc: "Pgsql-Performance@Postgresql. Org" , "Ray Ontko" Subject: Re: n_distinct way off, but following a pattern. In-reply-to: References: Comments: In-reply-to "Nick Fankhauser" message dated "Fri, 14 Nov 2003 16:27:16 -0500" Date: Fri, 14 Nov 2003 17:12:27 -0500 Message-ID: <5809.1068847947@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/125 X-Sequence-Number: 4685 "Nick Fankhauser" writes: > I'm seeing estimates for n_distinct that are way off for a large table Estimating n_distinct from a small sample is inherently a hard problem. I'm not surprised that the estimates would get better as the sample size increases. But maybe we can do better. The method we are currently using is this: /*---------- * Estimate the number of distinct values using the estimator * proposed by Haas and Stokes in IBM Research Report RJ 10025: * n*d / (n - f1 + f1*n/N) * where f1 is the number of distinct values that occurred * exactly once in our sample of n rows (from a total of N), * and d is the total number of distinct values in the sample. * This is their Duj1 estimator; the other estimators they * recommend are considerably more complex, and are numerically * very unstable when n is much smaller than N. It would be interesting to see exactly what inputs are going into this equation. Do you feel like adding some debug printouts into this code? Or just looking at the variables with a debugger? In 7.3 it's about line 1060 in src/backend/commands/analyze.c. BTW, this is already our second try at this problem, the original 7.2 equation didn't last long at all ... regards, tom lane From pgsql-performance-owner@postgresql.org Fri Nov 14 21:21:36 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EEFC8D1B4E1 for ; Sat, 15 Nov 2003 01:21:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 02155-08 for ; Fri, 14 Nov 2003 21:21:07 -0400 (AST) Received: from ALPHA9.ITS.MONASH.EDU.AU (alpha9.its.monash.edu.au [130.194.1.9]) by svr1.postgresql.org (Postfix) with ESMTP id E09D4D1B4F2 for ; Fri, 14 Nov 2003 21:21:03 -0400 (AST) Received: from localhost ([130.194.13.84]) by vaxh.its.monash.edu.au (PMDF V5.2-31 #39306) with ESMTP id <01L326TBENQ68ZRL0J@vaxh.its.monash.edu.au> for pgsql-performance@postgresql.org; Sat, 15 Nov 2003 12:20:21 +1100 Received: from blammo.its.monash.edu.au (localhost.its.monash.edu.au [127.0.0.1]) by localhost (Postfix) with ESMTP id 8C93239C003; Sat, 15 Nov 2003 12:20:21 +1100 (EST) Received: from bruce.csse.monash.edu.au (bruce.csse.monash.edu.au [130.194.64.3]) by blammo.its.monash.edu.au (Postfix) with ESMTP id 7E2162DC002; Sat, 15 Nov 2003 12:20:21 +1100 (EST) Received: from bruce.csse.monash.edu.au (localhost [127.0.0.1]) by bruce.csse.monash.edu.au (8.12.8+Sun/8.12.8) with ESMTP id hAF1KK8x013015; Sat, 15 Nov 2003 12:20:20 +1100 (EST) Received: from localhost (sgaric@localhost) by bruce.csse.monash.edu.au (8.12.8+Sun/8.12.2/Submit) with ESMTP id hAF1KKKr013012; Sat, 15 Nov 2003 12:20:20 +1100 (EST) Date: Sat, 15 Nov 2003 12:20:20 +1100 (EST) From: Slavisa Garic Subject: Re: INSERT extremely slow with large data sets (fwd) In-reply-to: <20031114220238.87764.qmail@web80210.mail.yahoo.com> To: George Essig Cc: Slavisa Garic , pgsql-performance@postgresql.org Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII Content-transfer-encoding: 7BIT X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/126 X-Sequence-Number: 4686 Does VACUUM ANALYZE help with the analysis or it also speeds up the process. I know i could try that before I ask but experiment is running now and I am too curious to wait :), Anyway thanks for the hint, Slavisa On Fri, 14 Nov 2003, George Essig wrote: > Slavisa Garic wrote: > > > Hi Everyone, > > > I am using PostgreSQL 7.3.2 and have used earlier versions (7.1.x > > onwards) > > and with all of them I noticed same problem with INSERTs when there is > > a > > large data set. Just to so you guys can compare time it takes to insert > > one row into a table when there are only few rows present and when > > there > > are thousands: > > Try running VACUUM ANALYZE periodically during inserts. I found this to help. > > George Essig > > From pgsql-performance-owner@postgresql.org Fri Nov 14 23:37:14 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C2A2CD1B4ED for ; Sat, 15 Nov 2003 03:37:10 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 31805-05 for ; Fri, 14 Nov 2003 23:36:43 -0400 (AST) Received: from relay01.kbs.net.au (relay01.kbs.net.au [203.220.32.149]) by svr1.postgresql.org (Postfix) with ESMTP id E90B5D1B4EB for ; Fri, 14 Nov 2003 23:36:38 -0400 (AST) Received: from [203.221.246.150] (helo=familyhealth.com.au) by relay01.kbs.net.au with esmtp (Exim 3.36 #1) id 1AKrEO-0007bJ-00; Sat, 15 Nov 2003 14:36:17 +1100 Message-ID: <3FB59F12.2090704@familyhealth.com.au> Date: Sat, 15 Nov 2003 11:35:46 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Neil Conway Cc: Josh Berkus , PostgreSQL Performance Subject: Re: Query question References: <3FB303E9.20903@familyhealth.com.au> <200311122120.28882.josh@agliodbs.com> <87llqjj5n0.fsf@mailbox.samurai.com> In-Reply-To: <87llqjj5n0.fsf@mailbox.samurai.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/127 X-Sequence-Number: 4687 >>The only thing you're adding to the query is a second SORT step, so it >>shouldn't require any more time/memory than the query's first SORT >>did. > > > Interesting -- I wonder if it would be possible for the optimizer to > detect this and avoid the redundant inner sort ... (/me muses to > himself) That's somethign I've wondered myself as well. Also - I wonder if the optimiser could be made smart enough to push down the outer LIMIT and OFFSET clauses into the subquery. Chris From pgsql-performance-owner@postgresql.org Sat Nov 15 06:21:49 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A9BD1D1B507 for ; Sat, 15 Nov 2003 10:21:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 94381-05 for ; Sat, 15 Nov 2003 06:21:19 -0400 (AST) Received: from zigo.dhs.org (as2-4-3.an.g.bonet.se [194.236.34.191]) by svr1.postgresql.org (Postfix) with ESMTP id BBA1FD1B502 for ; Sat, 15 Nov 2003 06:21:13 -0400 (AST) Received: from zigo.dhs.org (zigo [127.0.0.1]) by zigo.dhs.org (8.12.8/8.12.8) with ESMTP id hAFAL9L2013906; Sat, 15 Nov 2003 11:21:09 +0100 Received: from localhost (db@localhost) by zigo.dhs.org (8.12.8/8.12.8/Submit) with ESMTP id hAFAL9MU013902; Sat, 15 Nov 2003 11:21:09 +0100 Date: Sat, 15 Nov 2003 11:21:09 +0100 (CET) From: Dennis Bjorklund To: Tom Lane Cc: pgsql-performance@postgresql.org Subject: Re: Help with count(*) In-Reply-To: <4987.1068842326@sss.pgh.pa.us> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/128 X-Sequence-Number: 4688 On Fri, 14 Nov 2003, Tom Lane wrote: > I believe the previous discussion also agreed that we wanted to postpone > the freezing of now(), which currently also happens at BEGIN rather than > the first command after BEGIN. Or should that happen at the first call to now()? /me should ge back and try to find this previous discussion. -- /Dennis From pgsql-performance-owner@postgresql.org Sat Nov 15 09:14:09 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 49D28D1B50E for ; Sat, 15 Nov 2003 13:14:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 12835-08 for ; Sat, 15 Nov 2003 09:13:38 -0400 (AST) Received: from web80208.mail.yahoo.com (web80208.mail.yahoo.com [66.218.79.43]) by svr1.postgresql.org (Postfix) with SMTP id 36692D1B51A for ; Sat, 15 Nov 2003 09:13:35 -0400 (AST) Message-ID: <20031115131338.90329.qmail@web80208.mail.yahoo.com> Received: from [65.69.56.225] by web80208.mail.yahoo.com via HTTP; Sat, 15 Nov 2003 05:13:38 PST Date: Sat, 15 Nov 2003 05:13:38 -0800 (PST) From: George Essig Subject: Re: INSERT extremely slow with large data sets (fwd) To: Slavisa Garic Cc: Slavisa Garic , pgsql-performance@postgresql.org In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/129 X-Sequence-Number: 4689 --- Slavisa Garic wrote: > Does VACUUM ANALYZE help with the analysis or it also speeds up the > process. I know i could try that before I ask but experiment is running > now and I am too curious to wait :), > > Anyway thanks for the hint, > Slavisa > VACUUM ANALYZE will reclaim disk space and update statistics used by the optimizer to help execute queries faster. This could speed up your inserts. See http://www.postgresql.org/docs/7.3/static/sql-vacuum.html. George Essig From pgsql-novice-owner@postgresql.org Sat Nov 15 10:53:18 2003 X-Original-To: pgsql-novice-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1174AD1B51A; Sat, 15 Nov 2003 14:53:16 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38798-01; Sat, 15 Nov 2003 10:52:48 -0400 (AST) Received: from smtp1.nodak.edu (smtp1.NoDak.edu [134.129.111.50]) by svr1.postgresql.org (Postfix) with ESMTP id BF86BD1B4FC; Sat, 15 Nov 2003 10:52:42 -0400 (AST) Received: from ndsu.nodak.edu (webmail1.ndsu.NoDak.edu [134.129.111.141]) by smtp1.nodak.edu (8.11.6/8.11.6) with SMTP id hAFEqkB24187; Sat, 15 Nov 2003 08:52:46 -0600 Received: from 134.129.110.64 (SquirrelMail authenticated user radha.manohar) by webmail.ndsu.nodak.edu with HTTP; Sat, 15 Nov 2003 08:52:45 -0600 (CST) Message-ID: <3152.134.129.110.64.1068907965.squirrel@webmail.ndsu.nodak.edu> Date: Sat, 15 Nov 2003 08:52:45 -0600 (CST) Subject: Re: Error in transaction processing From: To: , X-Priority: 3 Importance: Normal X-Mailer: SquirrelMail (version 1.2.11) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.2 tagged_above=0.0 required=5.0 tests=BAYES_30, NO_REAL_NAME X-Spam-Level: X-Archive-Number: 200311/108 X-Sequence-Number: 8811 When I execute a transaction using embedded sql statements in a c program, I get the error, Error in transaction processing. I could see from the documentation that it means, "Postgres signalled to us that we cannot start, commit or rollback the transaction" I don't find any mistakes in the transaction statements. What can I do to correct this error? Your response would be very much appreciated. Thanks and Regards, Radha From pgsql-performance-owner@postgresql.org Sat Nov 15 16:21:25 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1EEE4D1C92F for ; Sat, 15 Nov 2003 20:21:25 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 83894-09 for ; Sat, 15 Nov 2003 16:20:54 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id 252F1D1C93C for ; Sat, 15 Nov 2003 16:20:53 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 7D7A61EB0; Sat, 15 Nov 2003 15:20:45 -0500 (EST) To: Tom Lane Cc: Will LaShell , Hannu Krosing , Christopher Browne , pgsql-performance@postgresql.org Subject: Re: Help with count(*) From: Neil Conway In-Reply-To: <4987.1068842326@sss.pgh.pa.us> (Tom Lane's message of "Fri, 14 Nov 2003 15:38:46 -0500") References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> Date: Sat, 15 Nov 2003 15:20:43 -0500 Message-ID: <87isllbcys.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/131 X-Sequence-Number: 4691 Tom Lane writes: > (I believe the previous discussion also agreed that we wanted to > postpone the freezing of now(), which currently also happens at > BEGIN rather than the first command after BEGIN.) That doesn't make sense to me: from a user's perspective, the "start of the transaction" is when the BEGIN is issued, regardless of any tricks we may play in the backend. Making now() return the time the current transaction started is reasonably logical; making now() return "the time when the first command after the BEGIN in the current transaction was issued" makes a lot less sense to me. -Neil From pgsql-hackers-owner@postgresql.org Sun Nov 16 07:58:56 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5FEC6D1D20B for ; Sun, 16 Nov 2003 11:58:54 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 26209-09 for ; Sun, 16 Nov 2003 07:58:26 -0400 (AST) Received: from fuji.krosing.net (217-159-136-226-dsl.kt.estpak.ee [217.159.136.226]) by svr1.postgresql.org (Postfix) with ESMTP id 7F79CD1E1A5 for ; Sun, 16 Nov 2003 07:58:22 -0400 (AST) Received: from fuji.krosing.net (localhost.localdomain [127.0.0.1]) by fuji.krosing.net (8.12.8/8.12.8) with ESMTP id hAGBwDJC017704; Sun, 16 Nov 2003 13:58:13 +0200 Received: (from hannu@localhost) by fuji.krosing.net (8.12.8/8.12.8/Submit) id hAGBwAO2017702; Sun, 16 Nov 2003 13:58:10 +0200 X-Authentication-Warning: fuji.krosing.net: hannu set sender to hannu@tm.ee using -f Subject: start of transaction (was: Re: [PERFORM] Help with count(*)) From: Hannu Krosing To: Neil Conway Cc: Tom Lane , Will LaShell , Christopher Browne , "pgsql-hackers@postgresql.org" In-Reply-To: <87isllbcys.fsf@mailbox.samurai.com> References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1068983887.17671.10.camel@fuji.krosing.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Sun, 16 Nov 2003 13:58:08 +0200 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/838 X-Sequence-Number: 47126 Redirected to -hackers Neil Conway kirjutas L, 15.11.2003 kell 22:20: > Tom Lane writes: > > (I believe the previous discussion also agreed that we wanted to > > postpone the freezing of now(), which currently also happens at > > BEGIN rather than the first command after BEGIN.) > > That doesn't make sense to me: from a user's perspective, the "start > of the transaction" is when the BEGIN is issued, regardless of any > tricks we may play in the backend. For me, the "start of transaction" is not about time, but about grouping a set of statements into one. So making the exact moment of "start" be the first statement that actually does something with data seems perfectly reasonable. If you really need to preserve time, do "select current_timestamp" and use the result. > Making now() return the time the current transaction started is > reasonably logical; making now() return "the time when the first > command after the BEGIN in the current transaction was issued" makes a > lot less sense to me. for me "the time the current transactuion is started" == "the time when the first command after the BEGIN in the current transaction was issued" and thus I see no conflict here ;) Delaying the locking effects of transactions as long as possible can increase performance overall, not just for pathological clients that sit on idle open transactions. Probably the latest time we can start the transaction is ath the start of executor step after the first statement in a transaction is planned and optimized. --------------- Hannu From pgsql-performance-owner@postgresql.org Sun Nov 16 09:53:23 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C1F81D1C952 for ; Sun, 16 Nov 2003 13:53:17 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 51875-02 for ; Sun, 16 Nov 2003 09:52:50 -0400 (AST) Received: from scaup.mail.pas.earthlink.net (scaup.mail.pas.earthlink.net [207.217.120.49]) by svr1.postgresql.org (Postfix) with ESMTP id BCD36D1B536 for ; Sun, 16 Nov 2003 09:52:46 -0400 (AST) Received: from 0-1pool116-139.nas15.indianapolis1.in.us.da.qwest.net ([65.128.116.139] helo=nick) by scaup.mail.pas.earthlink.net with smtp (Exim 3.33 #1) id 1ALNKV-0000R4-00; Sun, 16 Nov 2003 05:52:44 -0800 Reply-To: From: "Nick Fankhauser" To: "Tom Lane" Cc: "Pgsql-Performance@Postgresql. Org" , "Ray Ontko" Subject: Re: n_distinct way off, but following a pattern. Date: Sun, 16 Nov 2003 08:52:29 -0500 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) In-Reply-To: <5809.1068847947@sss.pgh.pa.us> Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/132 X-Sequence-Number: 4692 > It would be interesting to see exactly what inputs are going into this > equation. Do you feel like adding some debug printouts into this code? > Or just looking at the variables with a debugger? In 7.3 it's about > line 1060 in src/backend/commands/analyze.c. Tom- I don't really have time to follow up at this moment, but I think this would be interesting to look into, so I'll plan to dig into it over the Thanksgiving Holiday when I'll have a little time free to follow up on some fun projects. Your pointers should let me get into it pretty quickly. In the meantime, I'll just set up a cron job that runs behind my nightly analyze to put the correct numbers into pg_statistic on the tables that this affects. Thanks- -Nick From pgsql-hackers-owner@postgresql.org Sun Nov 16 10:50:58 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 66033D1B563 for ; Sun, 16 Nov 2003 14:50:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 59708-05 for ; Sun, 16 Nov 2003 10:50:24 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 35F88D1D20B for ; Sun, 16 Nov 2003 10:50:15 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAGEoB19026135; Sun, 16 Nov 2003 09:50:11 -0500 (EST) To: Hannu Krosing Cc: Neil Conway , Will LaShell , Christopher Browne , "pgsql-hackers@postgresql.org" Subject: Re: start of transaction (was: Re: [PERFORM] Help with count(*)) In-reply-to: <1068983887.17671.10.camel@fuji.krosing.net> References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> <1068983887.17671.10.camel@fuji.krosing.net> Comments: In-reply-to Hannu Krosing message dated "Sun, 16 Nov 2003 13:58:08 +0200" Date: Sun, 16 Nov 2003 09:50:11 -0500 Message-ID: <26134.1068994211@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/840 X-Sequence-Number: 47128 Hannu Krosing writes: > Probably the latest time we can start the transaction is ath the start > of executor step after the first statement in a transaction is planned > and optimized. The transaction has to exist before it can take locks, so the above would not fly. A complete example of what we have to think about is: BEGIN; SET TRANSACTION ISOLATION LEVEL SERIALIZABLE; LOCK TABLE foo; UPDATE foo ... -- or in general a SELECT/UPDATE/INSERT/DELETE query ... etc ... The transaction snapshot *must* be set at the time of the first query (here, the UPDATE). It obviously can't be later, and it cannot be earlier either, because in this sort of example you need the requested locks to be taken before the snapshot is set. The transaction must be created (as observed by other backends, in particular VACUUM) not later than the LOCK statement, else there is nothing that can own the lock. In principle though, the effects of BEGIN and perhaps SET could be strictly local to the current backend, and only when we hit a LOCK or query do we create the transaction externally. In practice the problem we observe is clients that issue BEGIN and then go to sleep (typically because of poorly-designed autocommit behavior in interface libraries). Postponing externally-visible creation of the transaction to the first command after BEGIN would be enough to get around the real-world issues, and it would not require code changes nearly as extensive as trying to let other stuff like SET happen "before" the transaction starts. There isn't any compelling implementation reason when to freeze the value of now(). Reasonable options are 1. at BEGIN (current behavior) 2. at transaction's external creation 3. at freezing of transaction snapshot #1 and #2 are actually the same at the moment, but could be decoupled as sketched above, in which case the behavior of #2 would effectively become "at first command afte BEGIN". In the previous thread: http://archives.postgresql.org/pgsql-hackers/2003-03/msg01178.php I argued that now() should be frozen at the time of the transaction snapshot, and I still think that that's a defensible behavior. regards, tom lane From pgsql-performance-owner@postgresql.org Sun Nov 16 11:23:16 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D0CA6D1E1A3 for ; Sun, 16 Nov 2003 15:23:13 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 61560-04 for ; Sun, 16 Nov 2003 11:22:46 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id BE0C6D1E19C for ; Sun, 16 Nov 2003 11:22:41 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAGFMh19026339; Sun, 16 Nov 2003 10:22:44 -0500 (EST) To: Neil Conway Cc: Will LaShell , Hannu Krosing , Christopher Browne , pgsql-performance@postgresql.org Subject: Re: Help with count(*) In-reply-to: <87isllbcys.fsf@mailbox.samurai.com> References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> Comments: In-reply-to Neil Conway message dated "Sat, 15 Nov 2003 15:20:43 -0500" Date: Sun, 16 Nov 2003 10:22:43 -0500 Message-ID: <26338.1068996163@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/133 X-Sequence-Number: 4693 Neil Conway writes: > Tom Lane writes: >> (I believe the previous discussion also agreed that we wanted to >> postpone the freezing of now(), which currently also happens at >> BEGIN rather than the first command after BEGIN.) > That doesn't make sense to me: from a user's perspective, the "start > of the transaction" is when the BEGIN is issued, regardless of any > tricks we may play in the backend. That's defensible when the user issued the BEGIN himself. When the BEGIN is coming from some interface library's autocommit logic, it's a lot less defensible. If you consult the archives, you will find actual user complaints about "why is now() returning a very old time?" that we traced to use of interface layers that handle "commit()" by issuing "COMMIT; BEGIN;". When BEGIN actually is issued by live application logic, I'd expect it to be followed immediately by some kind of command --- so the user would be unable to tell the difference in practice. Hannu moved this thread to -hackers, please follow up there if you want to discuss it more. regards, tom lane From pgsql-hackers-owner@postgresql.org Sun Nov 16 11:52:31 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 92623D1D207 for ; Sun, 16 Nov 2003 15:52:26 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 69009-03 for ; Sun, 16 Nov 2003 11:51:59 -0400 (AST) Received: from zigo.dhs.org (as2-4-3.an.g.bonet.se [194.236.34.191]) by svr1.postgresql.org (Postfix) with ESMTP id C291CD1E1A8 for ; Sun, 16 Nov 2003 11:51:53 -0400 (AST) Received: from zigo.dhs.org (zigo [127.0.0.1]) by zigo.dhs.org (8.12.8/8.12.8) with ESMTP id hAGFpoL2012182; Sun, 16 Nov 2003 16:51:50 +0100 Received: from localhost (db@localhost) by zigo.dhs.org (8.12.8/8.12.8/Submit) with ESMTP id hAGFpnhl012178; Sun, 16 Nov 2003 16:51:49 +0100 Date: Sun, 16 Nov 2003 16:51:49 +0100 (CET) From: Dennis Bjorklund To: Tom Lane Cc: Hannu Krosing , Neil Conway , Will LaShell , Christopher Browne , "pgsql-hackers@postgresql.org" Subject: Re: start of transaction (was: Re: [PERFORM] Help with In-Reply-To: <26134.1068994211@sss.pgh.pa.us> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/841 X-Sequence-Number: 47129 On Sun, 16 Nov 2003, Tom Lane wrote: > There isn't any compelling implementation reason when to freeze the > value of now(). Reasonable options are > 1. at BEGIN (current behavior) > 2. at transaction's external creation > 3. at freezing of transaction snapshot > #1 and #2 are actually the same at the moment, but could be decoupled > as sketched above, in which case the behavior of #2 would effectively > become "at first command afte BEGIN". > > I argued that now() should be frozen at the time of the transaction > snapshot, and I still think that that's a defensible behavior. Is it important exactly what value is returned as long as it's the same in the whole transaction? I think not. To me it would be just as logical to fix it at the first call to now() in the transaction. The first time you call it you get the actual time as it is now and the next time you get the same as before since every operation in the transaction logically happens at the same time. If you don't call now() at all, the system time will not be fetched at all. -- /Dennis From pgsql-hackers-owner@postgresql.org Sun Nov 16 18:56:22 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D8BEED1E13E for ; Sun, 16 Nov 2003 22:56:17 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 26154-08 for ; Sun, 16 Nov 2003 18:55:49 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id BCF0ED1B555 for ; Sun, 16 Nov 2003 18:55:46 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 4FDF71E37; Sun, 16 Nov 2003 17:55:42 -0500 (EST) To: Hannu Krosing Cc: Tom Lane , Will LaShell , Christopher Browne , "pgsql-hackers@postgresql.org" Subject: Re: start of transaction (was: Re: [PERFORM] Help with count(*)) From: Neil Conway In-Reply-To: <1068983887.17671.10.camel@fuji.krosing.net> (Hannu Krosing's message of "Sun, 16 Nov 2003 13:58:08 +0200") References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> <1068983887.17671.10.camel@fuji.krosing.net> Date: Sun, 16 Nov 2003 17:55:41 -0500 Message-ID: <873ccnc49e.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/853 X-Sequence-Number: 47141 Hannu Krosing writes: > For me, the "start of transaction" is not about time, but about grouping > a set of statements into one. So making the exact moment of "start" be > the first statement that actually does something with data seems > perfectly reasonable. This might be a perfectly logical change in semantics, but what benefit does it provide over the old way of doing things? What does BEGIN actually do now, from a user's perspective? At present, it "starts a transaction block", which is pretty simple. If we adopted the proposed change, it would "change the state of the system so that the next command is part of a new transaction". This is naturally more complex; but more importantly, what benefit does it ACTUALLY provide to the user? (I can't see one, but perhaps I'm missing something...) > Delaying the locking effects of transactions as long as possible can > increase performance overall, not just for pathological clients that sit > on idle open transactions. I agree, but this is irrelevant to the semantics of now(). -Neil From pgsql-hackers-owner@postgresql.org Sun Nov 16 19:18:40 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2C682D1E13E for ; Sun, 16 Nov 2003 23:18:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 35486-03 for ; Sun, 16 Nov 2003 19:18:04 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id 13F2DD1E13B for ; Sun, 16 Nov 2003 19:18:02 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 418E51F73; Sun, 16 Nov 2003 18:18:03 -0500 (EST) To: Tom Lane Cc: Will LaShell , Hannu Krosing , Christopher Browne , pgsql-hackers@postgresql.org Subject: start of transaction (was: Re: [PERFORM] Help with count(*)) From: Neil Conway In-Reply-To: <26338.1068996163@sss.pgh.pa.us> (Tom Lane's message of "Sun, 16 Nov 2003 10:22:43 -0500") References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> <26338.1068996163@sss.pgh.pa.us> Date: Sun, 16 Nov 2003 18:18:02 -0500 Message-ID: <87n0avaonp.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/854 X-Sequence-Number: 47142 Tom Lane writes: > That's defensible when the user issued the BEGIN himself. When the > BEGIN is coming from some interface library's autocommit logic, it's > a lot less defensible. If you consult the archives, you will find > actual user complaints about "why is now() returning a very old time?" > that we traced to use of interface layers that handle "commit()" by > issuing "COMMIT; BEGIN;". Hmmm... I agree this behavior isn't ideal, although I can see the case for viewing this as a mistake by the application developer: they are assuming that they know exactly when transactions begin, which is not a feature provided by their language interface. They should be using current_timestamp, and/or changing their language interface's configuration. That said, I think this is a minor irritation at best. The dual drawbacks of breaking backward compatibility and making the BEGIN semantics more confusing is enough to leave me satisfies with the status quo. If we do change this, I think Dennis' idea of making now() always return the same value within a given transaction is interesting: that might be a way to fix this problem without confusing the semantics of BEGIN. -Neil From pgsql-hackers-owner@postgresql.org Sun Nov 16 20:09:07 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1C27ED1E33C for ; Mon, 17 Nov 2003 00:09:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 44500-06 for ; Sun, 16 Nov 2003 20:08:34 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id DCDBFD1E33B for ; Sun, 16 Nov 2003 20:08:29 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAH08P19003799; Sun, 16 Nov 2003 19:08:25 -0500 (EST) To: Neil Conway Cc: Will LaShell , Hannu Krosing , Christopher Browne , pgsql-hackers@postgresql.org Subject: Re: start of transaction (was: Re: [PERFORM] Help with count(*)) In-reply-to: <87n0avaonp.fsf@mailbox.samurai.com> References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> <26338.1068996163@sss.pgh.pa.us> <87n0avaonp.fsf@mailbox.samurai.com> Comments: In-reply-to Neil Conway message dated "Sun, 16 Nov 2003 18:18:02 -0500" Date: Sun, 16 Nov 2003 19:08:24 -0500 Message-ID: <3798.1069027704@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/855 X-Sequence-Number: 47143 Neil Conway writes: > Hmmm... I agree this behavior isn't ideal, although I can see the case > for viewing this as a mistake by the application developer: they are > assuming that they know exactly when transactions begin, which is not > a feature provided by their language interface. Well, actually, it's a bug in the interface IMHO. But as I said in the last thread, it's a fairly widespread bug. We've been taking the position that the interface libraries should get fixed, and that's not happening. It's probably time to look at a server-side fix. > If we do change this, I think Dennis' idea of making now() always > return the same value within a given transaction is interesting: You mean the time of the first now() call? I thought that was an interesting idea also, but it's probably not going to look so hot when we complete the TODO item of adding access to the start-of-current-statement time. Having start-of-transaction be later than start-of-statement isn't gonna fly :-(. If we were willing to abandon that TODO item then I'd be interested in defining now() as Dennis suggested. regards, tom lane From pgsql-hackers-owner@postgresql.org Sun Nov 16 20:32:06 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id AF863D1E139 for ; Mon, 17 Nov 2003 00:32:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 52446-01 for ; Sun, 16 Nov 2003 20:31:31 -0400 (AST) Received: from candle.pha.pa.us (unknown [207.106.42.251]) by svr1.postgresql.org (Postfix) with ESMTP id 9B3E6D1B555 for ; Sun, 16 Nov 2003 20:31:27 -0400 (AST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.11.6) id hAH0VD328503; Sun, 16 Nov 2003 19:31:13 -0500 (EST) From: Bruce Momjian Message-Id: <200311170031.hAH0VD328503@candle.pha.pa.us> Subject: Re: start of transaction (was: Re: [PERFORM] Help with count(*)) In-Reply-To: <3798.1069027704@sss.pgh.pa.us> To: Tom Lane Date: Sun, 16 Nov 2003 19:31:13 -0500 (EST) Cc: Neil Conway , Will LaShell , Hannu Krosing , Christopher Browne , pgsql-hackers@postgresql.org X-Mailer: ELM [version 2.4ME+ PL108 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/856 X-Sequence-Number: 47144 Tom Lane wrote: > Neil Conway writes: > > Hmmm... I agree this behavior isn't ideal, although I can see the case > > for viewing this as a mistake by the application developer: they are > > assuming that they know exactly when transactions begin, which is not > > a feature provided by their language interface. > > Well, actually, it's a bug in the interface IMHO. But as I said in the > last thread, it's a fairly widespread bug. We've been taking the > position that the interface libraries should get fixed, and that's not > happening. It's probably time to look at a server-side fix. > > > If we do change this, I think Dennis' idea of making now() always > > return the same value within a given transaction is interesting: > > You mean the time of the first now() call? I thought that was an > interesting idea also, but it's probably not going to look so hot > when we complete the TODO item of adding access to > the start-of-current-statement time. Having start-of-transaction be > later than start-of-statement isn't gonna fly :-(. If we were willing > to abandon that TODO item then I'd be interested in defining now() as > Dennis suggested. Defining now() as the first call seems pretty arbitrary to me. I can't think of any time-based interface that has that API. And what if a trigger called now() in an earlier query and you didn't even know about it. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073 From pgsql-hackers-owner@postgresql.org Mon Nov 17 02:12:32 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D8DD7D1B57F for ; Mon, 17 Nov 2003 06:12:30 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 03646-04 for ; Mon, 17 Nov 2003 02:11:59 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id D741AD1B53A for ; Mon, 17 Nov 2003 02:11:58 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id A1DAB36E39; Mon, 17 Nov 2003 01:11:53 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1ALcc5-0000p5-00; Mon, 17 Nov 2003 01:11:53 -0500 To: pgsql-hackers@postgresql.org Subject: Re: start of transaction (was: Re: [PERFORM] Help with count(*)) References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> <1068983887.17671.10.camel@fuji.krosing.net> <873ccnc49e.fsf@mailbox.samurai.com> In-Reply-To: <873ccnc49e.fsf@mailbox.samurai.com> From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 17 Nov 2003 01:11:53 -0500 Message-ID: <87he13a5hy.fsf@stark.dyndns.tv> Lines: 29 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/858 X-Sequence-Number: 47146 Neil Conway writes: > What does BEGIN actually do now, from a user's perspective? I think you're thinking about this all wrong. BEGIN doesn't "do" anything. It's not a procedural statement, it's a declaration. It declares that the block of statements form a transaction so reads should be consistent and failures should be handled in a particular way to preserve data integrity. Given that declaration and the guarantees it requires of the database it's then up to the database to figure out what constraints that imposes on what the database can do and still meet the guarantees the BEGIN declaration requires. The more clever the database is about minimizing those restrictions the better as it means the database can run more efficiently. For what it's worth, this is how Oracle handles things too. On the command-line issuing a BEGIN following a COMMIT is just noise; you're _always_ in a transaction. A COMMIT ends the previous the transaction and implicitly starts the next transaction. But the snapshot isn't frozen until you first read from a table. I'm not sure what other databases do, but I think this is why clients behave like this. They think of BEGIN as a declaration and therefore initiating a COMMIT;BEGIN; at the end of every request is perfectly logical, and works fine in at least Oracle, and probably other databases. -- greg From pgsql-hackers-owner@postgresql.org Mon Nov 17 03:09:37 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 11CA3D1B95D for ; Mon, 17 Nov 2003 07:09:36 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 04323-05 for ; Mon, 17 Nov 2003 03:09:06 -0400 (AST) Received: from megazone.bigpanda.com (megazone.bigpanda.com [64.147.171.210]) by svr1.postgresql.org (Postfix) with ESMTP id 4D644D1B53A for ; Mon, 17 Nov 2003 03:09:05 -0400 (AST) Received: by megazone.bigpanda.com (Postfix, from userid 1001) id 728923532D; Sun, 16 Nov 2003 23:09:05 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by megazone.bigpanda.com (Postfix) with ESMTP id 710053532C; Sun, 16 Nov 2003 23:09:05 -0800 (PST) Date: Sun, 16 Nov 2003 23:09:05 -0800 (PST) From: Stephan Szabo To: Greg Stark Cc: pgsql-hackers@postgresql.org Subject: Re: start of transaction (was: Re: [PERFORM] Help with In-Reply-To: <87he13a5hy.fsf@stark.dyndns.tv> Message-ID: <20031116230227.H82963@megazone.bigpanda.com> References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> <1068983887.17671.10.camel@fuji.krosing.net> <873ccnc49e.fsf@mailbox.samurai.com> <87he13a5hy.fsf@stark.dyndns.tv> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/859 X-Sequence-Number: 47147 On Sun, 17 Nov 2003, Greg Stark wrote: > Neil Conway writes: > > > What does BEGIN actually do now, from a user's perspective? > > I think you're thinking about this all wrong. BEGIN doesn't "do" anything. > It's not a procedural statement, it's a declaration. It declares that the > block of statements form a transaction so reads should be consistent and > failures should be handled in a particular way to preserve data integrity. > > Given that declaration and the guarantees it requires of the database it's > then up to the database to figure out what constraints that imposes on what > the database can do and still meet the guarantees the BEGIN declaration > requires. The more clever the database is about minimizing those restrictions > the better as it means the database can run more efficiently. > > For what it's worth, this is how Oracle handles things too. On the > command-line issuing a BEGIN following a COMMIT is just noise; you're _always_ > in a transaction. A COMMIT ends the previous the transaction and implicitly > starts the next transaction. But the snapshot isn't frozen until you first > read from a table. The earlier portion of the described behavior is AFAICS not complient to SQL99 at least. COMMIT (without AND CHAIN) terminates a transaction and does not begin a new one. The new transaction does not begin until a transaction initiating command (for example START TRANSACTION, CREATE TABLE, INSERT, ...) is executed. The set of things you can do that aren't initiating is fairly small admittedly, but it's not a null set. From pgsql-hackers-owner@postgresql.org Mon Nov 17 06:17:10 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EDA0BD1B579 for ; Mon, 17 Nov 2003 10:17:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 30431-08 for ; Mon, 17 Nov 2003 06:16:38 -0400 (AST) Received: from fuji.krosing.net (217-159-136-226-dsl.kt.estpak.ee [217.159.136.226]) by svr1.postgresql.org (Postfix) with ESMTP id 775F0D1B584 for ; Mon, 17 Nov 2003 06:16:19 -0400 (AST) Received: from fuji.krosing.net (localhost.localdomain [127.0.0.1]) by fuji.krosing.net (8.12.8/8.12.8) with ESMTP id hAHAGDJC020262; Mon, 17 Nov 2003 12:16:13 +0200 Received: (from hannu@localhost) by fuji.krosing.net (8.12.8/8.12.8/Submit) id hAHAG6M9020260; Mon, 17 Nov 2003 12:16:06 +0200 X-Authentication-Warning: fuji.krosing.net: hannu set sender to hannu@tm.ee using -f Subject: Re: start of transaction (was: Re: [PERFORM] Help with From: Hannu Krosing To: Tom Lane Cc: Neil Conway , Will LaShell , Christopher Browne , "pgsql-hackers@postgresql.org" In-Reply-To: <3798.1069027704@sss.pgh.pa.us> References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> <26338.1068996163@sss.pgh.pa.us> <87n0avaonp.fsf@mailbox.samurai.com> <3798.1069027704@sss.pgh.pa.us> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1069064165.20092.20.camel@fuji.krosing.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Mon, 17 Nov 2003 12:16:06 +0200 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/862 X-Sequence-Number: 47150 Tom Lane kirjutas E, 17.11.2003 kell 02:08: > Neil Conway writes: > > Hmmm... I agree this behavior isn't ideal, although I can see the case > > for viewing this as a mistake by the application developer: they are > > assuming that they know exactly when transactions begin, which is not > > a feature provided by their language interface. > > Well, actually, it's a bug in the interface IMHO. But as I said in the > last thread, it's a fairly widespread bug. I'm not sure that it is a client-side bug. For example Oracle seems to _always_ have a transaction going, i.e. you can't be "outside" of transaction, and you use just COMMIT to commit old _and_start_new_ transaction. IIRC the same is true for DB2. For these database the BEGIN TRANSACTION command is mainly used for starting nested transactions, which we don't have. > We've been taking the > position that the interface libraries should get fixed, and that's not > happening. It's probably time to look at a server-side fix. Maybe "fixing" the interface libraries would make them incompatible with *DBC's for all other databases in some subtle ways ? ----------------- Hannu From pgsql-hackers-owner@postgresql.org Mon Nov 17 06:20:20 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 01C36D1B584 for ; Mon, 17 Nov 2003 10:20:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 43067-01 for ; Mon, 17 Nov 2003 06:19:50 -0400 (AST) Received: from fuji.krosing.net (217-159-136-226-dsl.kt.estpak.ee [217.159.136.226]) by svr1.postgresql.org (Postfix) with ESMTP id 41F7CD1B536 for ; Mon, 17 Nov 2003 06:19:47 -0400 (AST) Received: from fuji.krosing.net (localhost.localdomain [127.0.0.1]) by fuji.krosing.net (8.12.8/8.12.8) with ESMTP id hAHAJJJC020269; Mon, 17 Nov 2003 12:19:19 +0200 Received: (from hannu@localhost) by fuji.krosing.net (8.12.8/8.12.8/Submit) id hAHAJ6lF020266; Mon, 17 Nov 2003 12:19:06 +0200 X-Authentication-Warning: fuji.krosing.net: hannu set sender to hannu@tm.ee using -f Subject: Re: start of transaction (was: Re: [PERFORM] Help with From: Hannu Krosing To: Bruce Momjian Cc: Tom Lane , Neil Conway , Will LaShell , Christopher Browne , "pgsql-hackers@postgresql.org" In-Reply-To: <200311170031.hAH0VD328503@candle.pha.pa.us> References: <200311170031.hAH0VD328503@candle.pha.pa.us> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1069064346.20092.27.camel@fuji.krosing.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Mon, 17 Nov 2003 12:19:06 +0200 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/863 X-Sequence-Number: 47151 Bruce Momjian kirjutas E, 17.11.2003 kell 02:31: > Defining now() as the first call seems pretty arbitrary to me. I can't > think of any time-based interface that has that API. And what if a > trigger called now() in an earlier query and you didn't even know about > it. That would be OK. The whole point of that previous discussion was to have now() that returns the same value over the span of the whole transaction. It would be even better to have now() that returns the time current transaction is COMMITted as this is the time other backend become aware of it ;) ----------- Hannu From pgsql-hackers-owner@postgresql.org Tue Nov 18 16:16:08 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 7E48DD1B584 for ; Mon, 17 Nov 2003 10:44:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 37760-05 for ; Mon, 17 Nov 2003 06:43:54 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 8EE16D1B536 for ; Mon, 17 Nov 2003 06:43:46 -0400 (AST) Received: from persistent.co.in shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Mon, 17 Nov 2003 03:43:46 -0700 Message-ID: <3FB8A650.9030603@persistent.co.in> Date: Mon, 17 Nov 2003 16:13:28 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "pgsql-hackers@postgresql.org" Subject: Re: start of transaction References: <200311141251.38786.mallah@trade-india.com> <1068831807.2540.44.camel@fuji.krosing.net> <3FB52FA7.8060508@lashell.net> <4987.1068842326@sss.pgh.pa.us> <87isllbcys.fsf@mailbox.samurai.com> <26338.1068996163@sss.pgh.pa.us> <87n0avaonp.fsf@mailbox.samurai.com> <3798.1069027704@sss.pgh.pa.us> <1069064165.20092.20.camel@fuji.krosing.net> In-Reply-To: <1069064165.20092.20.camel@fuji.krosing.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/1006 X-Sequence-Number: 47294 Hannu Krosing wrote: > Tom Lane kirjutas E, 17.11.2003 kell 02:08: > >>Neil Conway writes: >> >>>Hmmm... I agree this behavior isn't ideal, although I can see the case >>>for viewing this as a mistake by the application developer: they are >>>assuming that they know exactly when transactions begin, which is not >>>a feature provided by their language interface. >> >>Well, actually, it's a bug in the interface IMHO. But as I said in the >>last thread, it's a fairly widespread bug. > > > I'm not sure that it is a client-side bug. For example Oracle seems to > _always_ have a transaction going, i.e. you can't be "outside" of > transaction, and you use just COMMIT to commit old _and_start_new_ > transaction. > > IIRC the same is true for DB2. Actually, in oracle a new transaction starts with first DDL after a commit. That does not include DML BTW. And Damn.. Actually I recently fixed a "bug" where I had to force a start of transaction in Pro*C, immediately after commit. Otherwise a real start of transaction could be anywhere down the line, causing some weird concurrency issues. Rather than fiddling with oracle support, I would hack my source code, especially this is not the first oracle bug I have worked around....:-( The fact that I couldn't control exact transaction start was such a irritation to put it mildly.. I sooooo missed 'exec sql begin work' in ecpg..:-) >>We've been taking the >>position that the interface libraries should get fixed, and that's not >>happening. It's probably time to look at a server-side fix. I hope that does not compramise transaction control I have with libpq/ecpg etc. And when we are talking about interface libraries, how many of them are within PG control and how many are not? With languages maintenend by postgresql group, it should behave correctly, right? E.g pl/perl,pl/python etc. And for other interface libraries, what are they exactly? php? Can't we just send them a stinker/patch to get that damn thing right(Whatever wrong they are doing. I have kinda lost thread on it..:-) Was it exact time of transaction start v/s now()?) Shridhar From pgsql-performance-owner@postgresql.org Mon Nov 17 10:43:13 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 88DFFD1DABB for ; Mon, 17 Nov 2003 14:43:11 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 82740-04 for ; Mon, 17 Nov 2003 10:42:44 -0400 (AST) Received: from mail.libertyrms.com (unknown [209.167.124.227]) by svr1.postgresql.org (Postfix) with ESMTP id 2D07AD1D8AE for ; Mon, 17 Nov 2003 10:42:40 -0400 (AST) Received: from [10.1.2.130] (helo=dba2) by mail.libertyrms.com with esmtp (Exim 3.22 #3 (Debian)) id 1ALkaS-0001jz-00 for ; Mon, 17 Nov 2003 09:42:44 -0500 Received: by dba2 (Postfix, from userid 1019) id 0063ACBB8; Mon, 17 Nov 2003 09:42:43 -0500 (EST) Date: Mon, 17 Nov 2003 09:42:43 -0500 From: Andrew Sullivan To: pgsql-performance@postgresql.org Subject: Re: INSERT extremely slow with large data sets (fwd) Message-ID: <20031117144243.GE1716@libertyrms.info> Mail-Followup-To: Andrew Sullivan , pgsql-performance@postgresql.org References: <20031115131338.90329.qmail@web80208.mail.yahoo.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20031115131338.90329.qmail@web80208.mail.yahoo.com> User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/134 X-Sequence-Number: 4694 On Sat, Nov 15, 2003 at 05:13:38AM -0800, George Essig wrote: > > VACUUM ANALYZE will reclaim disk space and update statistics used Strictly speaking, it does not reclaim disk space. It merely marks it as available, assuming you have enough room in your free space map. VACUUM FULL reclaims disk space, i.e. it compacts the data files and returns that space to the operating system. A -- ---- Andrew Sullivan 204-4141 Yonge Street Afilias Canada Toronto, Ontario Canada M2P 2A8 +1 416 646 3304 x110 From pgsql-performance-owner@postgresql.org Mon Nov 17 10:52:03 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 89F56D1B536 for ; Mon, 17 Nov 2003 14:51:58 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 81735-07 for ; Mon, 17 Nov 2003 10:51:31 -0400 (AST) Received: from jacobson.be (c20753a.mael.bostream.se [217.215.37.134]) by svr1.postgresql.org (Postfix) with SMTP id 70E78D1B50C for ; Mon, 17 Nov 2003 10:51:26 -0400 (AST) Received: (qmail 2433 invoked by uid 99); 17 Nov 2003 14:51:21 -0000 Received: from dhcp-5.localdomain (dhcp-5.localdomain [192.168.0.5]) by mail.jacobson.be (IMP) with HTTP for ; Mon, 17 Nov 2003 15:51:20 +0100 Message-ID: <1069080680.3fb8e068eb941@mail.jacobson.be> Date: Mon, 17 Nov 2003 15:51:20 +0100 From: Joel Jacobson To: pgsql-performance@postgresql.org Subject: Backup/restore of pg_statistics MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.2.2 X-Originating-IP: 192.168.0.5 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/135 X-Sequence-Number: 4695 Hi, I understand that it is not possible to occasionally re-plan the queries in a PL/pgSQL function without dropping and re-creating the function. I think it would be useful if the queries in a PL/pgSQL function could be re-planned on-the-fly. When a lot of data has been added/modified and ANALYZE is suitable to run, it would also be a great idea to re-plan the queries used in PL/pgSQL functions. I understand that this is not possible? The only way would be to DROP/CREATE the functions or to use EXECUTE. I don't think EXECUTE is an option, because preparing the queries every time the function is called is in my case not necessary and just a waste of performance. As a work-around, I am forced to, 1. populate the database with a lot of test data, 2. run ANALYZE, 3. and finally, create the PL/pgSQL functions The prepared queries in the functions will now be sufficiently optimized. I don't think this is a nice solution. I also thought of a slightly better solution, but I don't know if it is possible. My idea was to populate the database once and then save the data in pg_statistics generated by ANALYZE to a file. Every time the database needs to be created, the statistics could then be restored thus making the planner produce "future-optimized" queries when the PL/pgSQL functions are created, even though the database is empty. I would greatly appreciate any help/comments. Thank you. Joel Jacobson From pgsql-performance-owner@postgresql.org Mon Nov 17 12:39:17 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5743FD1B537 for ; Mon, 17 Nov 2003 16:39:15 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 04418-09 for ; Mon, 17 Nov 2003 12:38:43 -0400 (AST) Received: from martin.sysdetect.com (martin.sysdetect.com [65.209.102.1]) by svr1.postgresql.org (Postfix) with ESMTP id A9500D1B54C for ; Mon, 17 Nov 2003 12:38:42 -0400 (AST) Received: (from mail@localhost) by martin.sysdetect.com (8.11.4/8.11.3) id hAHGcg402482 for ; Mon, 17 Nov 2003 16:38:42 GMT Received: from winwood.sysdetect.com(172.16.1.1) via SMTP by mail.sysdetect.com, id smtpdA26397; Mon Nov 17 16:38:38 2003 Received: from sysd.com (thriller.sysdetect.com [172.16.1.33]) by winwood.sysdetect.com (8.11.6/8.11.6) with ESMTP id hAHGccb19973 for ; Mon, 17 Nov 2003 11:38:38 -0500 Message-ID: <3FB8F98D.1010707@sysd.com> Date: Mon, 17 Nov 2003 11:38:37 -0500 From: Rich Cullingford User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030703 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Top n queries and GROUP BY Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/136 X-Sequence-Number: 4696 All, This is a straight SQL question, maybe not appropriate for a performance list, but... I have a simple stock holdings setup: => select * from t1; nam | co | num -----+-----------+------ joe | ibm | 600 abe | ibm | 1500 joe | cisco | 1200 abe | cisco | 800 joe | novell | 500 joe | microsoft | 200 What I would like to see is a Top-n-holdings-by-name", e.g, for n=2: nam | co | num ----------+--------+----- joe | cisco | 1200 joe | ibm | 600 abe | ibm | 1500 abe | cisco | 800 I can get part of the way by using a LIMIT clause in a subquery, e.g, => select 'abe', a.co, a.num from (select co, num from t1 where nam='abe' order by num desc limit 2) as a; ?column? | co | num ----------+-------+------ abe | ibm | 1500 abe | cisco | 800 but I can't figure out a correlated subquery (or GROUP BY arrangement or anything else) that will cycle through the names. I vaguely remember that these kinds or queries are hard to do in standard SQL, but I was hoping that PG, with its extensions... Thanks, Rich Cullingford rculling@sysd.com From pgsql-performance-owner@postgresql.org Mon Nov 17 13:57:00 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id F11C5D1B50A for ; Mon, 17 Nov 2003 17:56:58 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 17384-02 for ; Mon, 17 Nov 2003 13:56:27 -0400 (AST) Received: from martin.sysdetect.com (martin.sysdetect.com [65.209.102.1]) by svr1.postgresql.org (Postfix) with ESMTP id 5A757D1B8B9 for ; Mon, 17 Nov 2003 13:56:26 -0400 (AST) Received: (from mail@localhost) by martin.sysdetect.com (8.11.4/8.11.3) id hAHHuQ105024 for ; Mon, 17 Nov 2003 17:56:26 GMT Received: from winwood.sysdetect.com(172.16.1.1) via SMTP by mail.sysdetect.com, id smtpdl10528; Mon Nov 17 17:56:23 2003 Received: from sysd.com (thriller.sysdetect.com [172.16.1.33]) by winwood.sysdetect.com (8.11.6/8.11.6) with ESMTP id hAHHuNb22743; Mon, 17 Nov 2003 12:56:23 -0500 Message-ID: <3FB90BC7.5090602@sysd.com> Date: Mon, 17 Nov 2003 12:56:23 -0500 From: Rich Cullingford User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030703 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Rich Cullingford Cc: pgsql-performance@postgresql.org Subject: Re: Top n queries and GROUP BY References: <3FB8F98D.1010707@sysd.com> In-Reply-To: <3FB8F98D.1010707@sysd.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/137 X-Sequence-Number: 4697 Rich Cullingford wrote: > All, > This is a straight SQL question, maybe not appropriate for a performance > list, but... > > I have a simple stock holdings setup: > > => select * from t1; > nam | co | num > -----+-----------+------ > joe | ibm | 600 > abe | ibm | 1500 > joe | cisco | 1200 > abe | cisco | 800 > joe | novell | 500 > joe | microsoft | 200 > > What I would like to see is a Top-n-holdings-by-name", e.g, for n=2: > > nam | co | num > ----------+--------+----- > joe | cisco | 1200 > joe | ibm | 600 > abe | ibm | 1500 > abe | cisco | 800 > > I can get part of the way by using a LIMIT clause in a subquery, e.g, > > => select 'abe', a.co, a.num from (select co, num from t1 where > nam='abe' order by num desc limit 2) as a; > ?column? | co | num > ----------+-------+------ > abe | ibm | 1500 > abe | cisco | 800 > > but I can't figure out a correlated subquery (or GROUP BY arrangement or > anything else) that will cycle through the names. I vaguely remember > that these kinds or queries are hard to do in standard SQL, but I was > hoping that PG, with its extensions... I forgot about row subqueries; for n=3, for example: => SELECT * FROM t1 WHERE (nam,co,num) IN (SELECT nam,co,num FROM t1 b where b.nam=t1.nam order by num desc limit 3) order by nam, num desc; nam | co | num -----+--------+------ abe | ibm | 1500 abe | cisco | 800 joe | cisco | 1200 joe | ibm | 600 joe | novell | 500 (5 rows) Seems to work... Thanks all, Rich Cullingford rculling@sysd.com From pgsql-performance-owner@postgresql.org Mon Nov 17 18:22:09 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 23C1CD1C975 for ; Mon, 17 Nov 2003 22:22:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 65292-09 for ; Mon, 17 Nov 2003 18:21:38 -0400 (AST) Received: from nitrogen.id.pl (nitrogen.id.pl [193.178.214.5]) by svr1.postgresql.org (Postfix) with SMTP id 4D5DDD1CA06 for ; Mon, 17 Nov 2003 18:21:14 -0400 (AST) Received: (qmail 8003 invoked by uid 0); 17 Nov 2003 22:21:00 -0000 Received: from unknown (HELO siaco.id.pl) (213.25.114.8) by smtp.id.pl with SMTP; 17 Nov 2003 22:21:00 -0000 Received: (qmail 4818 invoked by uid 1000); 17 Nov 2003 22:20:59 -0000 Date: Mon, 17 Nov 2003 23:20:58 +0100 From: Ryszard Lach To: pgsql-performance@postgresql.org Subject: duration logging setting in 7.4 Message-ID: <20031117222058.GD3248@siaco.id.pl> Reply-To: Ryszard Lach Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-2 Content-Disposition: inline X-My-GPG-Key: echo | mail -s "send key pub" ryszard@lach.name User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/138 X-Sequence-Number: 4698 Hi. I'm trying to set run-time environment in pgsql7.4 so, that it prints all statements with duration time, but I can't understand why setting log_min_duration_statement to '0' causes printing to syslog plenty of lines ending with 'duration: statement:', i.e. without any statement string (except expected ones). Can anybody help me? Richard. -- "First they ignore you. Then they laugh at you. Then they fight you. Then you win." - Mohandas Gandhi. From pgsql-hackers-owner@postgresql.org Mon Nov 17 20:53:37 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8C515D1C9B7 for ; Tue, 18 Nov 2003 00:53:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 91875-04 for ; Mon, 17 Nov 2003 20:53:05 -0400 (AST) Received: from candle.pha.pa.us (unknown [207.106.42.251]) by svr1.postgresql.org (Postfix) with ESMTP id E3CB7D1B527 for ; Mon, 17 Nov 2003 20:53:00 -0400 (AST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.11.6) id hAI0qpw10716; Mon, 17 Nov 2003 19:52:51 -0500 (EST) From: Bruce Momjian Message-Id: <200311180052.hAI0qpw10716@candle.pha.pa.us> Subject: Re: start of transaction (was: Re: [PERFORM] Help with count(*)) In-Reply-To: <1069064346.20092.27.camel@fuji.krosing.net> To: Hannu Krosing Date: Mon, 17 Nov 2003 19:52:51 -0500 (EST) Cc: Tom Lane , Neil Conway , Will LaShell , Christopher Browne , "pgsql-hackers@postgresql.org" X-Mailer: ELM [version 2.4ME+ PL108 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/896 X-Sequence-Number: 47184 Hannu Krosing wrote: > Bruce Momjian kirjutas E, 17.11.2003 kell 02:31: > > > Defining now() as the first call seems pretty arbitrary to me. I can't > > think of any time-based interface that has that API. And what if a > > trigger called now() in an earlier query and you didn't even know about > > it. > > That would be OK. The whole point of that previous discussion was to > have now() that returns the same value over the span of the whole > transaction. I think my issue is that there isn't any predictable way for a user to know when the now() time is recorded. By using start of transaction, at least we know for sure the point in time it is showing. > It would be even better to have now() that returns the time current > transaction is COMMITted as this is the time other backend become aware > of it ;) True, but implementing that would be very hard. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073 From pgsql-performance-owner@postgresql.org Mon Nov 17 22:38:50 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EC552D1CA63 for ; Tue, 18 Nov 2003 02:38:44 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 14682-04 for ; Mon, 17 Nov 2003 22:38:17 -0400 (AST) Received: from candle.pha.pa.us (unknown [207.106.42.251]) by svr1.postgresql.org (Postfix) with ESMTP id 109B6D1CAD1 for ; Mon, 17 Nov 2003 22:37:05 -0400 (AST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.11.6) id hAI2b7s29417; Mon, 17 Nov 2003 21:37:07 -0500 (EST) From: Bruce Momjian Message-Id: <200311180237.hAI2b7s29417@candle.pha.pa.us> Subject: Re: duration logging setting in 7.4 In-Reply-To: <20031117222058.GD3248@siaco.id.pl> To: Ryszard Lach Date: Mon, 17 Nov 2003 21:37:07 -0500 (EST) Cc: pgsql-performance@postgresql.org X-Mailer: ELM [version 2.4ME+ PL108 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/139 X-Sequence-Number: 4699 Ryszard Lach wrote: > > Hi. > > I'm trying to set run-time environment in pgsql7.4 so, that it prints > all statements with duration time, but I can't understand why setting > log_min_duration_statement to '0' causes printing to syslog plenty of > lines ending with 'duration: statement:', i.e. without any statement > string (except expected ones). Can anybody help me? Can you show us some of the log file? If I do: test=> set log_min_duration_statement = 0; SET test=> select 1; ?column? ---------- 1 (1 row) I get: LOG: duration: 0.861 ms statement: select 1; -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073 From pgsql-hackers-owner@postgresql.org Tue Nov 18 03:35:43 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 01AD4D1B575 for ; Tue, 18 Nov 2003 07:35:42 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 72114-08 for ; Tue, 18 Nov 2003 03:35:11 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 9EB2BD1B50F for ; Tue, 18 Nov 2003 03:35:09 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAI7Yq19017647; Tue, 18 Nov 2003 02:34:53 -0500 (EST) To: Bruce Momjian Cc: Hannu Krosing , Neil Conway , Will LaShell , Christopher Browne , "pgsql-hackers@postgresql.org" Subject: Re: start of transaction (was: Re: [PERFORM] Help with count(*)) In-reply-to: <200311180052.hAI0qpw10716@candle.pha.pa.us> References: <200311180052.hAI0qpw10716@candle.pha.pa.us> Comments: In-reply-to Bruce Momjian message dated "Mon, 17 Nov 2003 19:52:51 -0500" Date: Tue, 18 Nov 2003 02:34:52 -0500 Message-ID: <17646.1069140892@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/952 X-Sequence-Number: 47240 Bruce Momjian writes: > Hannu Krosing wrote: >> It would be even better to have now() that returns the time current >> transaction is COMMITted as this is the time other backend become aware >> of it ;) > True, but implementing that would be very hard. Son, that was a *joke* ... regards, tom lane From pgsql-performance-owner@postgresql.org Tue Nov 18 05:18:42 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E7427D1D80C for ; Tue, 18 Nov 2003 09:18:39 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 86718-08 for ; Tue, 18 Nov 2003 05:18:09 -0400 (AST) Received: from nitrogen.id.pl (nitrogen.id.pl [193.178.214.5]) by svr1.postgresql.org (Postfix) with SMTP id 3617AD1B4E8 for ; Tue, 18 Nov 2003 05:18:04 -0400 (AST) Received: (qmail 7469 invoked by uid 0); 18 Nov 2003 09:16:50 -0000 Received: from unknown (HELO siaco.id.pl) (213.25.114.8) by smtp.id.pl with SMTP; 18 Nov 2003 09:16:50 -0000 Received: (qmail 8615 invoked by uid 1000); 18 Nov 2003 09:16:46 -0000 Date: Tue, 18 Nov 2003 10:16:46 +0100 From: Ryszard Lach To: pgsql-performance@postgresql.org Cc: Bruce Momjian Subject: Re: duration logging setting in 7.4 Message-ID: <20031118091646.GA8223@siaco.id.pl> Reply-To: Ryszard Lach References: <20031117222058.GD3248@siaco.id.pl> <200311180237.hAI2b7s29417@candle.pha.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-2 Content-Disposition: inline In-Reply-To: <200311180237.hAI2b7s29417@candle.pha.pa.us> X-My-GPG-Key: echo | mail -s "send key pub" ryszard@lach.name User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/140 X-Sequence-Number: 4700 On Mon, Nov 17, 2003 at 09:37:07PM -0500, Bruce Momjian wrote: > Ryszard Lach wrote: > > > > Hi. > > > > I'm trying to set run-time environment in pgsql7.4 so, that it prints > > all statements with duration time, but I can't understand why setting > > log_min_duration_statement to '0' causes printing to syslog plenty of > > lines ending with 'duration: statement:', i.e. without any statement > > string (except expected ones). Can anybody help me? > > Can you show us some of the log file? If I do: > Sure. Nov 18 10:05:20 postgres[1348]: [318-1] LOG: duration: 0.297 ms statement: Nov 18 10:05:20 postgres[1311]: [5477-1] LOG: duration: 0.617 ms statement: Nov 18 10:05:20 postgres[1312]: [5134-1] LOG: duration: 0.477 ms statement: Nov 18 10:05:20 postgres[1349]: [318-1] LOG: duration: 0.215 ms statement: Nov 18 10:05:20 postgres[1313]: [5449-1] LOG: duration: 0.512 ms statement: Nov 18 10:05:20 postgres[1314]: [5534-1] LOG: duration: 0.420 ms statement: Nov 18 10:05:20 postgres[1330]: [772-1] LOG: duration: 1.386 ms statement: SELECT * FROM mytablemius WHERE id = 0; Nov 18 10:05:20 postgres[1315]: [5757-1] LOG: duration: 0.417 ms statement: Nov 18 10:05:20 postgres[1316]: [5885-1] LOG: duration: 0.315 ms statement: Nov 18 10:05:20 postgres[1317]: [5914-1] LOG: duration: 0.301 ms statement: Nov 18 10:05:20 postgres[1318]: [5990-1] LOG: duration: 0.293 ms statement: Nov 18 10:05:20 postgres[1319]: [6009-1] LOG: duration: 0.211 ms statement: Nov 18 10:05:20 postgres[1320]: [6039-1] LOG: duration: 0.188 ms statement: this is with setting log_duration = false log_statement = false log_min_duration_statement = 0 The amount of lines containing statement string is nearly the same ase before upgrade (from 7.3), all other lines are extra. I don't know if this can be a reason, this is on a pretty busy machine (ca. 100 selects/second, but loadavg lower then 0.9), I'm logging postgres through syslog. Richard. -- "First they ignore you. Then they laugh at you. Then they fight you. Then you win." - Mohandas Gandhi. From pgsql-performance-owner@postgresql.org Tue Nov 18 06:10:28 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 39DA7D1C9AB for ; Tue, 18 Nov 2003 10:10:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 96375-06 for ; Tue, 18 Nov 2003 06:09:54 -0400 (AST) Received: from stan.aopsys.com (m173.net81-64-15.noos.fr [81.64.15.173]) by svr1.postgresql.org (Postfix) with ESMTP id 58943D1B515 for ; Tue, 18 Nov 2003 06:09:51 -0400 (AST) Received: from localhost ([127.0.0.1] helo=news.nerim.net) by stan.aopsys.com with esmtp (Exim 3.36 #1 (Debian)) id 1AM2g8-00063z-00 for ; Tue, 18 Nov 2003 11:01:48 +0100 To: Subject: Join on incompatible types References: From: Laurent Martelli Date: Tue, 18 Nov 2003 11:01:48 +0100 In-Reply-To: (scott marlowe's message of "Wed, 29 Oct 2003 14:51:18 -0700 (MST)") Message-ID: <87k75ym1v7.fsf@news.nerim.net> User-Agent: Gnus/5.1002 (Gnus v5.10.2) Emacs/20.7 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/141 X-Sequence-Number: 4701 >>>>> "scott" == scott marlowe writes: [...] scott> Note here: scott> Merge Join (cost=1788.68..4735.71 rows=1 width=85) (actual scott> time=597.540..1340.526 rows=20153 loops=1) Merge Cond: scott> ("outer".id = "inner".id) scott> This estimate is WAY off. Are both of those fields indexed scott> and analyzed? Have you tried upping the statistics target on scott> those two fields? I assume they are compatible types. Should I understand that a join on incompatible types (such as integer and varchar) may lead to bad performances ? -- Laurent Martelli laurent@aopsys.com Java Aspect Components http://www.aopsys.com/ http://jac.aopsys.com From pgsql-performance-owner@postgresql.org Tue Nov 18 06:46:26 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A56E9D1D803 for ; Tue, 18 Nov 2003 10:46:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 03224-06 for ; Tue, 18 Nov 2003 06:45:55 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 3C673D1C9DA for ; Tue, 18 Nov 2003 06:45:52 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Tue, 18 Nov 2003 03:45:49 -0700 Message-ID: <3FB9F855.7010905@myrealbox.com> Date: Tue, 18 Nov 2003 16:15:41 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Laurent Martelli Cc: pgsql-performance@postgresql.org Subject: Re: Join on incompatible types References: <87k75ym1v7.fsf@news.nerim.net> In-Reply-To: <87k75ym1v7.fsf@news.nerim.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/142 X-Sequence-Number: 4702 Laurent Martelli wrote: >>>>>>"scott" == scott marlowe writes: > > > [...] > > scott> Note here: > > scott> Merge Join (cost=1788.68..4735.71 rows=1 width=85) (actual > scott> time=597.540..1340.526 rows=20153 loops=1) Merge Cond: > scott> ("outer".id = "inner".id) > > scott> This estimate is WAY off. Are both of those fields indexed > scott> and analyzed? Have you tried upping the statistics target on > scott> those two fields? I assume they are compatible types. > > Should I understand that a join on incompatible types (such as integer > and varchar) may lead to bad performances ? Conversely, you should enforce strict type compatibility in comparisons for getting any good plans..:-) Shridhar From pgsql-performance-owner@postgresql.org Tue Nov 18 15:47:41 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3037ED1CA92 for ; Tue, 18 Nov 2003 13:33:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 33208-01 for ; Tue, 18 Nov 2003 09:32:59 -0400 (AST) Received: from stan.aopsys.com (m173.net81-64-15.noos.fr [81.64.15.173]) by svr1.postgresql.org (Postfix) with ESMTP id CC052D1D880 for ; Tue, 18 Nov 2003 09:32:54 -0400 (AST) Received: from localhost ([127.0.0.1] helo=news.nerim.net) by stan.aopsys.com with esmtp (Exim 3.36 #1 (Debian)) id 1AM5qe-0006H6-00; Tue, 18 Nov 2003 14:24:52 +0100 To: Shridhar Daithankar Cc: pgsql-performance@postgresql.org Subject: Re: Join on incompatible types References: <87k75ym1v7.fsf@news.nerim.net> <3FB9F855.7010905@myrealbox.com> From: Laurent Martelli Date: Tue, 18 Nov 2003 14:24:51 +0100 In-Reply-To: <3FB9F855.7010905@myrealbox.com> (Shridhar Daithankar's message of "Tue, 18 Nov 2003 16:15:41 +0530") Message-ID: <87oev9lsgs.fsf@news.nerim.net> User-Agent: Gnus/5.1002 (Gnus v5.10.2) Emacs/20.7 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/145 X-Sequence-Number: 4705 >>>>> "Shridhar" == Shridhar Daithankar writes: Shridhar> Laurent Martelli wrote: [...] >> Should I understand that a join on incompatible types (such as >> integer and varchar) may lead to bad performances ? Shridhar> Conversely, you should enforce strict type compatibility Shridhar> in comparisons for getting any good plans..:-) Ha ha, now I understand why a query of mine was so sluggish. Is there a chance I could achieve the good perfs without having he same types ? I've tried a CAST in the query, but it's even a little worse than without it. However, using a view to cast integers into varchar gives acceptable results (see at the end). I'm using Postgresql 7.3.4. iprofil-jac=# EXPLAIN ANALYZE SELECT * from classes where exists (select value from lists where lists.id='16' and lists.value=classes.id); QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------- Seq Scan on classes (cost=0.00..5480289.75 rows=9610 width=25) (actual time=31.68..7321.56 rows=146 loops=1) Filter: (subplan) SubPlan -> Index Scan using lists_id on lists (cost=0.00..285.12 rows=1 width=8) (actual time=0.38..0.38 rows=0 loops=19220) Index Cond: (id = 16) Filter: ((value)::text = ($0)::text) Total runtime: 7321.72 msec iprofil-jac=# EXPLAIN ANALYZE SELECT * from classes2 where exists (select value from lists where lists.id='16' and lists.value=classes2.id); QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------- Seq Scan on classes2 (cost=0.00..5923.87 rows=500 width=64) (actual time=0.76..148.20 rows=146 loops=1) Filter: (subplan) SubPlan -> Index Scan using lists_value on lists (cost=0.00..5.90 rows=1 width=8) (actual time=0.01..0.01 rows=0 loops=19220) Index Cond: ((id = 16) AND (value = $0)) Total runtime: 148.34 msec -- -- Tables classes and classes2 are populated with the same data, they -- only differ on the type of the "id" column. -- iprofil-jac=# \d classes Table "public.classes" Colonne | Type | Modifications ---------+-------------------+--------------- id | integer | not null classid | character varying | Index: classes_pkey primary key btree (id) iprofil-jac=# \d classes2 Table "public.classes2" Colonne | Type | Modifications ---------+-------------------+--------------- id | character varying | not null classid | character varying | Index: classes2_pkey primary key btree (id) iprofil-jac=# \d lists Table "public.lists" Colonne | Type | Modifications ---------+-------------------+--------------- id | integer | not null index | integer | not null value | character varying | Index: lists_index unique btree (id, "index"), lists_id btree (id), lists_value btree (id, value) -- -- IT'S EVEN BETTER WITH A JOIN -- iprofil-jac=# EXPLAIN ANALYZE SELECT * from lists join classes on classes.id=lists.value where lists.id='16'; QUERY PLAN ------------------------------------------------------------------------------------------------------------------- Nested Loop (cost=0.00..90905.88 rows=298 width=41) (actual time=53.93..9327.87 rows=146 loops=1) Join Filter: (("inner".id)::text = ("outer".value)::text) -> Seq Scan on lists (cost=0.00..263.43 rows=146 width=16) (actual time=8.38..9.70 rows=146 loops=1) Filter: (id = 16) -> Seq Scan on classes (cost=0.00..333.20 rows=19220 width=25) (actual time=0.00..28.45 rows=19220 loops=146) Total runtime: 9328.35 msec iprofil-jac=# EXPLAIN ANALYZE SELECT * from lists join classes2 on classes2.id=lists.value where lists.id='16'; QUERY PLAN -------------------------------------------------------------------------------------------------------------------------------------- Merge Join (cost=268.67..324.09 rows=16 width=80) (actual time=9.59..65.55 rows=146 loops=1) Merge Cond: ("outer".id = "inner".value) -> Index Scan using classes2_pkey on classes2 (cost=0.00..52.00 rows=1000 width=64) (actual time=0.03..40.83 rows=18778 loops=1) -> Sort (cost=268.67..269.03 rows=146 width=16) (actual time=9.50..9.56 rows=146 loops=1) Sort Key: lists.value -> Seq Scan on lists (cost=0.00..263.43 rows=146 width=16) (actual time=8.83..9.17 rows=146 loops=1) Filter: (id = 16) Total runtime: 65.73 msec -- -- CASTING IN THE QUERY IS NO GOOD -- iprofil-jac=# EXPLAIN ANALYZE SELECT * from lists join classes on CAST(classes.id AS character varying)=lists.value where lists.id='16'; QUERY PLAN ------------------------------------------------------------------------------------------------------------------- Nested Loop (cost=0.00..90905.88 rows=298 width=41) (actual time=69.03..10017.26 rows=146 loops=1) Join Filter: ((("inner".id)::text)::character varying = "outer".value) -> Seq Scan on lists (cost=0.00..263.43 rows=146 width=16) (actual time=20.64..22.03 rows=146 loops=1) Filter: (id = 16) -> Seq Scan on classes (cost=0.00..333.20 rows=19220 width=25) (actual time=0.00..30.45 rows=19220 loops=146) Total runtime: 10017.72 msec -- -- CREATING A VIEW IS BETTER -- iprofil-jac=# CREATE VIEW classes3 as SELECT CAST(id AS varchar), classid from classes; iprofil-jac=# EXPLAIN ANALYZE SELECT * from classes3 where exists (select value from lists where lists.id='16' and lists.value=classes3.id); QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------- Seq Scan on classes (cost=0.00..113853.60 rows=9610 width=25) (actual time=0.91..192.31 rows=146 loops=1) Filter: (subplan) SubPlan -> Index Scan using lists_value on lists (cost=0.00..5.91 rows=1 width=8) (actual time=0.01..0.01 rows=0 loops=19220) Index Cond: ((id = 16) AND (value = (($0)::text)::character varying)) Total runtime: 192.47 msec -- Laurent Martelli laurent@aopsys.com Java Aspect Components http://www.aopsys.com/ http://jac.aopsys.com From pgsql-performance-owner@postgresql.org Tue Nov 18 11:08:26 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 74776D1CA01 for ; Tue, 18 Nov 2003 15:08:23 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45258-08 for ; Tue, 18 Nov 2003 11:07:56 -0400 (AST) Received: from candle.pha.pa.us (unknown [207.106.42.251]) by svr1.postgresql.org (Postfix) with ESMTP id 5B9AED1B4EF for ; Tue, 18 Nov 2003 11:07:51 -0400 (AST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.11.6) id hAIF7mX04577; Tue, 18 Nov 2003 10:07:48 -0500 (EST) From: Bruce Momjian Message-Id: <200311181507.hAIF7mX04577@candle.pha.pa.us> Subject: Re: duration logging setting in 7.4 In-Reply-To: <20031118091646.GA8223@siaco.id.pl> To: Ryszard Lach Date: Tue, 18 Nov 2003 10:07:48 -0500 (EST) Cc: pgsql-performance@postgresql.org X-Mailer: ELM [version 2.4ME+ PL108 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/143 X-Sequence-Number: 4703 Wow, that is strange. If you don't use syslog, do you see the proper output? If you turn on log_statement, do you see the statements? --------------------------------------------------------------------------- Ryszard Lach wrote: > On Mon, Nov 17, 2003 at 09:37:07PM -0500, Bruce Momjian wrote: > > Ryszard Lach wrote: > > > > > > Hi. > > > > > > I'm trying to set run-time environment in pgsql7.4 so, that it prints > > > all statements with duration time, but I can't understand why setting > > > log_min_duration_statement to '0' causes printing to syslog plenty of > > > lines ending with 'duration: statement:', i.e. without any statement > > > string (except expected ones). Can anybody help me? > > > > Can you show us some of the log file? If I do: > > > > Sure. > > Nov 18 10:05:20 postgres[1348]: [318-1] LOG: duration: 0.297 ms statement: > Nov 18 10:05:20 postgres[1311]: [5477-1] LOG: duration: 0.617 ms statement: > Nov 18 10:05:20 postgres[1312]: [5134-1] LOG: duration: 0.477 ms statement: > Nov 18 10:05:20 postgres[1349]: [318-1] LOG: duration: 0.215 ms statement: > Nov 18 10:05:20 postgres[1313]: [5449-1] LOG: duration: 0.512 ms statement: > Nov 18 10:05:20 postgres[1314]: [5534-1] LOG: duration: 0.420 ms statement: > Nov 18 10:05:20 postgres[1330]: [772-1] LOG: duration: 1.386 ms statement: SELECT * FROM mytablemius WHERE id = 0; > Nov 18 10:05:20 postgres[1315]: [5757-1] LOG: duration: 0.417 ms statement: > Nov 18 10:05:20 postgres[1316]: [5885-1] LOG: duration: 0.315 ms statement: > Nov 18 10:05:20 postgres[1317]: [5914-1] LOG: duration: 0.301 ms statement: > Nov 18 10:05:20 postgres[1318]: [5990-1] LOG: duration: 0.293 ms statement: > Nov 18 10:05:20 postgres[1319]: [6009-1] LOG: duration: 0.211 ms statement: > Nov 18 10:05:20 postgres[1320]: [6039-1] LOG: duration: 0.188 ms statement: > > > this is with setting > > log_duration = false > log_statement = false > log_min_duration_statement = 0 > > The amount of lines containing statement string is nearly the same ase before > upgrade (from 7.3), all other lines are extra. > > I don't know if this can be a reason, this is on a pretty busy machine (ca. 100 > selects/second, but loadavg lower then 0.9), I'm logging postgres through syslog. > > Richard. > > -- > "First they ignore you. Then they laugh at you. Then they > fight you. Then you win." - Mohandas Gandhi. > -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073 From pgsql-performance-owner@postgresql.org Tue Nov 18 13:01:49 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6E94AD1C915 for ; Tue, 18 Nov 2003 17:01:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 64500-09 for ; Tue, 18 Nov 2003 13:01:17 -0400 (AST) Received: from main.gmane.org (main.gmane.org [80.91.224.249]) by svr1.postgresql.org (Postfix) with ESMTP id 41658D1B4E8 for ; Tue, 18 Nov 2003 13:01:16 -0400 (AST) Received: from list by main.gmane.org with local (Exim 3.35 #1 (Debian)) id 1AM9E2-0002q6-00 for ; Tue, 18 Nov 2003 18:01:14 +0100 X-Injected-Via-Gmane: http://gmane.org/ To: pgsql-performance@postgresql.org Received: from sea.gmane.org ([80.91.224.252]) by main.gmane.org with esmtp (Exim 3.35 #1 (Debian)) id 1AM9E1-0002py-00 for ; Tue, 18 Nov 2003 18:01:13 +0100 Received: from news by sea.gmane.org with local (Exim 3.35 #1 (Debian)) id 1AM9E1-0003ID-00 for ; Tue, 18 Nov 2003 18:01:13 +0100 From: Harald Fuchs Subject: Re: Top n queries and GROUP BY Date: 18 Nov 2003 18:01:13 +0100 Organization: Linux Private Site Lines: 50 Message-ID: References: <3FB8F98D.1010707@sysd.com> Reply-To: hf99@protecting.net Mime-Version: 1.0 (generated by tm-edit 7.90) Content-Type: text/plain; charset=US-ASCII X-Complaints-To: usenet@sea.gmane.org X-No-Archive: yes X-Newsreader: Gnus v5.7/Emacs 20.7 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/144 X-Sequence-Number: 4704 In article <3FB8F98D.1010707@sysd.com>, Rich Cullingford writes: > All, > This is a straight SQL question, maybe not appropriate for a performance > list, but... > I have a simple stock holdings setup: > => select * from t1; > nam | co | num > -----+-----------+------ > joe | ibm | 600 > abe | ibm | 1500 > joe | cisco | 1200 > abe | cisco | 800 > joe | novell | 500 > joe | microsoft | 200 > What I would like to see is a Top-n-holdings-by-name", e.g, for n=2: > nam | co | num > ----------+--------+----- > joe | cisco | 1200 > joe | ibm | 600 > abe | ibm | 1500 > abe | cisco | 800 > I can get part of the way by using a LIMIT clause in a subquery, e.g, > => select 'abe', a.co, a.num from (select co, num from t1 where > nam='abe' order by num desc limit 2) as a; > ?column? | co | num > ----------+-------+------ > abe | ibm | 1500 > abe | cisco | 800 > but I can't figure out a correlated subquery (or GROUP BY arrangement or > anything else) that will cycle through the names. I vaguely remember > that these kinds or queries are hard to do in standard SQL, but I was > hoping that PG, with its extensions... How about an outer join? SELECT x1.nam, x1.co, x1.num FROM t1 x1 LEFT JOIN t1 x2 ON x2.nam = x1.nam AND x2.num > x1.num GROUP BY x1.nam, x1.co, x1.num HAVING count(*) < 2 ORDER BY x1.nam, x1.num DESC From pgsql-performance-owner@postgresql.org Tue Nov 18 20:01:53 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6D5CDD1CA0F for ; Wed, 19 Nov 2003 00:01:49 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 42378-03 for ; Tue, 18 Nov 2003 20:01:21 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 4213DD1B520 for ; Tue, 18 Nov 2003 20:01:18 -0400 (AST) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3941320; Tue, 18 Nov 2003 16:02:05 -0800 Content-Type: text/plain; charset="us-ascii" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: Shridhar Daithankar Subject: More detail on settings for pgavd? Date: Tue, 18 Nov 2003 15:58:45 -0800 User-Agent: KMail/1.4.3 Cc: pgsql-performance@postgresql.org MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200311181558.45847.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/146 X-Sequence-Number: 4706 Shridhar, I was looking at the -V/-v and -A/-a settings in pgavd, and really don't=20 understand how the calculation works. According to the readme, if I set -= v=20 to 1000 and -V to 2 (the defaults) for a table with 10,000 rows, pgavd woul= d=20 only vacuum after 21,000 rows had been updated. This seems wrong. Can you clear this up a little? I'd like to tweak these settings but can'= t=20 without being better aquainted with the calculation. Also, you may want to reverse your default ratio for Vacuum/analyze frequen= cy.=20=20 True, analyze is a less expensive operation than Vacuum, but it's also need= ed=20 less often -- only when the *distribution* of data changes. I've seen=20 databases where the optimal vacuum/analyze frequency was every 10 min/once= =20 per day. --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Nov 18 21:20:49 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A2EAAD1D7BD for ; Wed, 19 Nov 2003 01:20:48 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 52268-09 for ; Tue, 18 Nov 2003 21:20:20 -0400 (AST) Received: from mail.viatornetworks.com (unknown [216.224.225.50]) by svr1.postgresql.org (Postfix) with ESMTP id 85822D1D674 for ; Tue, 18 Nov 2003 21:20:15 -0400 (AST) Received: from [10.1.0.6] (216-224-225-54.phxinternet.net [216.224.225.54] (may be forged)) by mail.viatornetworks.com (8.9.3/8.9.3) with ESMTP id SAA09099 for ; Tue, 18 Nov 2003 18:20:02 -0700 Subject: High Processor consumption From: Benjamin Bostow To: pgsql-performance@postgresql.org Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.8 (1.0.8-11) Date: 18 Nov 2003 18:19:14 -0700 Message-Id: <1069204784.1955.16.camel@localhost.localdomain> Mime-Version: 1.0 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/147 X-Sequence-Number: 4707 I am running RH 7.3 running Apache 1.3.27-2 and PostgreSQL 7.2.3-5.73. When having 100+ users connected to my server I notice that postmaster consumes up wards of 90% of the processor and I hardly am higher than 10% idle. I did notice that when I kill apache and postmaster that my idle processor percentage goes to 95% or higher. I am looking on ways that I can find what connections are making the database processes run so high. If this could also tell which program is accessing it, it would be helpful. I have look through the documents on performance tuning postgresql and have adjusted my memory with little effect. I have even routed all traffic away from the Apache server so no load is on apache. I do have C programs that run and access the database. Any help will be greatly appreciated. Thanks in advance, Ben From pgsql-performance-owner@postgresql.org Wed Nov 19 02:33:09 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CE4B2D1D5A4 for ; Wed, 19 Nov 2003 06:33:08 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 05764-06 for ; Wed, 19 Nov 2003 02:32:38 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 98256D1B53A for ; Wed, 19 Nov 2003 02:32:36 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Tue, 18 Nov 2003 23:32:28 -0700 Message-ID: <3FBB0E77.4020105@myrealbox.com> Date: Wed, 19 Nov 2003 12:02:23 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: josh@agliodbs.com Cc: pgsql-performance@postgresql.org Subject: Re: More detail on settings for pgavd? References: <200311181558.45847.josh@agliodbs.com> In-Reply-To: <200311181558.45847.josh@agliodbs.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/149 X-Sequence-Number: 4709 Josh Berkus wrote: > Shridhar, > > I was looking at the -V/-v and -A/-a settings in pgavd, and really don't > understand how the calculation works. According to the readme, if I set -v > to 1000 and -V to 2 (the defaults) for a table with 10,000 rows, pgavd would > only vacuum after 21,000 rows had been updated. This seems wrong. > > Can you clear this up a little? I'd like to tweak these settings but can't > without being better aquainted with the calculation. > > Also, you may want to reverse your default ratio for Vacuum/analyze frequency. > True, analyze is a less expensive operation than Vacuum, but it's also needed > less often -- only when the *distribution* of data changes. I've seen > databases where the optimal vacuum/analyze frequency was every 10 min/once > per day. Will look into it. Give me a day or so. I am planning couple of other patches as well. May be over week end. Is this urgent? Shridhar From pgsql-performance-owner@postgresql.org Wed Nov 19 02:34:52 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 87912D1B50A for ; Wed, 19 Nov 2003 06:34:50 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 13715-06 for ; Wed, 19 Nov 2003 02:34:20 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 5468BD1C4EE for ; Wed, 19 Nov 2003 02:34:19 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Tue, 18 Nov 2003 23:34:19 -0700 Message-ID: <3FBB0EE4.6010805@myrealbox.com> Date: Wed, 19 Nov 2003 12:04:12 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Benjamin Bostow Cc: pgsql-performance@postgresql.org Subject: Re: High Processor consumption References: <1069204784.1955.16.camel@localhost.localdomain> In-Reply-To: <1069204784.1955.16.camel@localhost.localdomain> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/150 X-Sequence-Number: 4710 Benjamin Bostow wrote: > I am running RH 7.3 running Apache 1.3.27-2 and PostgreSQL 7.2.3-5.73. > When having 100+ users connected to my server I notice that postmaster > consumes up wards of 90% of the processor and I hardly am higher than > 10% idle. I did notice that when I kill apache and postmaster that my > idle processor percentage goes to 95% or higher. I am looking on ways > that I can find what connections are making the database processes run > so high. If this could also tell which program is accessing it, it would > be helpful. I have look through the documents on performance tuning > postgresql and have adjusted my memory with little effect. I have even > routed all traffic away from the Apache server so no load is on apache. > I do have C programs that run and access the database. Any help will be > greatly appreciated. Routinely the CPU load for postgresql translates to too much low shared buffers setting for requirement. What are your postgresql.conf tunings? Could you please post them? Shridhar From pgsql-performance-owner@postgresql.org Wed Nov 19 02:54:44 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E3D8DD1D5AB for ; Wed, 19 Nov 2003 06:54:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 14518-03 for ; Wed, 19 Nov 2003 02:54:13 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id C8C3DD1B4E6 for ; Wed, 19 Nov 2003 02:54:11 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Tue, 18 Nov 2003 23:54:10 -0700 Message-ID: <3FBB138C.4090303@myrealbox.com> Date: Wed, 19 Nov 2003 12:24:04 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Laurent Martelli Cc: pgsql-performance@postgresql.org Subject: Re: Join on incompatible types References: <87k75ym1v7.fsf@news.nerim.net> <3FB9F855.7010905@myrealbox.com> <87oev9lsgs.fsf@news.nerim.net> In-Reply-To: <87oev9lsgs.fsf@news.nerim.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/151 X-Sequence-Number: 4711 Laurent Martelli wrote: >>>>>>"Shridhar" == Shridhar Daithankar writes: > > > Shridhar> Laurent Martelli wrote: > > [...] > > >> Should I understand that a join on incompatible types (such as > >> integer and varchar) may lead to bad performances ? > > Shridhar> Conversely, you should enforce strict type compatibility > Shridhar> in comparisons for getting any good plans..:-) > > Ha ha, now I understand why a query of mine was so sluggish. > > Is there a chance I could achieve the good perfs without having he > same types ? I've tried a CAST in the query, but it's even a little > worse than without it. However, using a view to cast integers into > varchar gives acceptable results (see at the end). > > I'm using Postgresql 7.3.4. I am stripping the analyze outputs and directly jumping to the end. Can you try following? 1. Make all fields integer in all the table. 2. Try following query EXPLAIN ANALYZE SELECT * from lists join classes on classes.id=lists.value where lists.id='16'::integer; How does it affect the runtime? Shridhar From pgsql-performance-owner@postgresql.org Wed Nov 19 05:18:56 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DCD24D1D612 for ; Wed, 19 Nov 2003 09:18:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 32767-07 for ; Wed, 19 Nov 2003 05:18:23 -0400 (AST) Received: from p15140235.pureserver.info (sql-info.de [217.160.166.164]) by svr1.postgresql.org (Postfix) with ESMTP id B5F48D1C50F for ; Wed, 19 Nov 2003 05:18:21 -0400 (AST) Received: by p15140235.pureserver.info (Postfix, from userid 639) id 77036203A; Wed, 19 Nov 2003 10:18:23 +0100 (CET) Received: from iwate.sql-info.de (pD9517DFE.dip.t-dialin.net [217.81.125.254]) (using SSLv3 with cipher RC4-MD5 (128/128 bits)) (No client certificate requested) by p15140235.pureserver.info (Postfix) with ESMTP id 5948CEA1 for ; Wed, 19 Nov 2003 10:18:22 +0100 (CET) From: Ian Barwick To: pgsql-performance@postgresql.org Subject: TEXT column and indexing Date: Wed, 19 Nov 2003 10:18:18 +0100 User-Agent: KMail/1.5.1 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311191018.18492.barwick@gmx.net> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/152 X-Sequence-Number: 4712 I have this table: db=> \d object_property_value Table "db.object_property_value" Column | Type | Modifiers -----------------------+------------------------+-------------------- obj_property_value_id | integer | not null default nextval(... obj_property_id | integer | not null value | text | Indexes: "object_property_value_pkey" primary key, btree (obj_property_value_id) "opv_obj_property_id_ix" btree (obj_property_id) "opv_v_ix" btree (substr(value, 1, 128)) Foreign-key constraints: "object_property_fkey" FOREIGN KEY (obj_property_id) REFERENCES object_property(obj_property_id) ON UPDATE CASCADE ON DELETE CASCADE (long lines edited for readability). The table contains about 250,000 records and will grow at regular intervals. The 'value' column contains text of various lengths. The table is VACUUMed and ANALYZEd regularly and waxed on Sunday mornings. Database encoding is Unicode. Server is 7.4RC1 or 7.4RC2 and will be 7.4 ASAP. I want to query this table to match a specific value along the lines of: SELECT obj_property_id FROM object_property_value opv WHERE opv.value = 'foo' There will only be a few (at the moment 2 or 3) rows exactly matching 'foo'. This query will only be performed with values containing less than around 100 characters, which account for ca. 10% of all rows in the table. The performance is of course lousy: db=> EXPLAIN db-> SELECT obj_property_id db-> FROM object_property_value opv db-> WHERE opv.value = 'foo'; QUERY PLAN ----------------------------------------------------------------------------- Seq Scan on object_property_value opv (cost=0.00..12258.26 rows=2 width=4) Filter: (value = 'foo'::text) (2 rows) However, if I create a VARCHAR field containing the first 128 characters of the text field and index that, an index scan is used: db=> EXPLAIN db-> SELECT obj_property_id db-> FROM object_property_value opv db-> WHERE opv.opv_vc = 'foo'; QUERY PLAN ------------------------------------------------------------------------------------------- Index Scan using opv_vc_ix on object_property_value opv (cost=0.00..6.84 rows=2 width=4) Index Cond: ((opv_vc)::text = 'foo'::text) The question is therefore: can I get an index to work on the TEXT column? It is currently indexed with: "opv_v_ix" btree (substr(value, 1, 128)) which doesn't appear to have any effect. I am probably missing something obvious though. I can live with maintaining an extra VARCHAR column but would like to keep the table as simple as possible. (For anyone wondering: yes, I can access the data using tsearch2 - via a different table in this case - but this is not always appropriate). Thanks for any hints. Ian Barwick barwick@gmx.net From pgsql-performance-owner@postgresql.org Wed Nov 19 06:01:22 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id F20DED1D729 for ; Wed, 19 Nov 2003 10:01:20 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45224-01 for ; Wed, 19 Nov 2003 06:00:52 -0400 (AST) Received: from stan.aopsys.com (m173.net81-64-15.noos.fr [81.64.15.173]) by svr1.postgresql.org (Postfix) with ESMTP id A9B82D1D630 for ; Wed, 19 Nov 2003 06:00:43 -0400 (AST) Received: from localhost ([127.0.0.1] helo=news.nerim.net) by stan.aopsys.com with esmtp (Exim 3.36 #1 (Debian)) id 1AMP0t-0007ZF-00; Wed, 19 Nov 2003 10:52:43 +0100 To: Shridhar Daithankar Cc: pgsql-performance@postgresql.org Subject: Re: Join on incompatible types References: <87k75ym1v7.fsf@news.nerim.net> <3FB9F855.7010905@myrealbox.com> <87oev9lsgs.fsf@news.nerim.net> <3FBB138C.4090303@myrealbox.com> From: Laurent Martelli Organization: AOPSYS Date: Wed, 19 Nov 2003 10:52:43 +0100 In-Reply-To: <3FBB138C.4090303@myrealbox.com> (Shridhar Daithankar's message of "Wed, 19 Nov 2003 12:24:04 +0530") Message-ID: <87u150k7mc.fsf@news.nerim.net> User-Agent: Gnus/5.1002 (Gnus v5.10.2) Emacs/20.7 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/153 X-Sequence-Number: 4713 >>>>> "Shridhar" == Shridhar Daithankar writes: Shridhar> Laurent Martelli wrote: >>>>>>> "Shridhar" == Shridhar Daithankar >>>>>>> writes: Shridhar> Laurent Martelli wrote: >> [...] >> Should I understand that a join on incompatible types >> (such as >> integer and varchar) may lead to bad performances ? Shridhar> Conversely, you should enforce strict type compatibility Shridhar> in comparisons for getting any good plans..:-) >> Ha ha, now I understand why a query of mine was so sluggish. Is >> there a chance I could achieve the good perfs without having he >> same types ? I've tried a CAST in the query, but it's even a >> little worse than without it. However, using a view to cast >> integers into varchar gives acceptable results (see at the end). >> I'm using Postgresql 7.3.4. Shridhar> I am stripping the analyze outputs and directly jumping to Shridhar> the end. Shridhar> Can you try following? Shridhar> 1. Make all fields integer in all the table. I can't do this because lists.values contains non integer data which do not refer to a classes.id value. It may sound weird. This is because it's a generic schema for a transparent persistence framework. The solution for me would rather be to have varchar everywhere. Shridhar> 2. Try following query EXPLAIN ANALYZE SELECT * from lists Shridhar> join classes on classes.id=lists.value where Shridhar> lists.id='16'::integer; Shridhar> How does it affect the runtime? Shridhar> Shridhar -- Laurent Martelli laurent@aopsys.com Java Aspect Components http://www.aopsys.com/ http://jac.aopsys.com From pgsql-performance-owner@postgresql.org Wed Nov 19 06:20:06 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 32AE9D1D5CF for ; Wed, 19 Nov 2003 10:20:04 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 40541-09 for ; Wed, 19 Nov 2003 06:19:35 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 86E61D1D80C for ; Wed, 19 Nov 2003 06:19:32 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Wed, 19 Nov 2003 03:19:30 -0700 Message-ID: <3FBB43AB.30302@myrealbox.com> Date: Wed, 19 Nov 2003 15:49:23 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Laurent Martelli Cc: pgsql-performance@postgresql.org Subject: Re: Join on incompatible types References: <87k75ym1v7.fsf@news.nerim.net> <3FB9F855.7010905@myrealbox.com> <87oev9lsgs.fsf@news.nerim.net> <3FBB138C.4090303@myrealbox.com> <87u150k7mc.fsf@news.nerim.net> In-Reply-To: <87u150k7mc.fsf@news.nerim.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/154 X-Sequence-Number: 4714 Laurent Martelli wrote: >>>>>>"Shridhar" == Shridhar Daithankar writes: > Shridhar> I am stripping the analyze outputs and directly jumping to > Shridhar> the end. > > Shridhar> Can you try following? > > Shridhar> 1. Make all fields integer in all the table. > > I can't do this because lists.values contains non integer data which > do not refer to a classes.id value. It may sound weird. This is > because it's a generic schema for a transparent persistence framework. Fine .I understand. So instead of using a field value, can you use integer version of that field? (Was that one of your queries used that? I deleted the OP) > The solution for me would rather be to have varchar everywhere. You need to cast every occurance of that varchar field appropriately, to start with. The performance might suffer as well for numbers. > Shridhar> 2. Try following query EXPLAIN ANALYZE SELECT * from lists > Shridhar> join classes on classes.id=lists.value where > Shridhar> lists.id='16'::integer; classes.id=lists.value::integer. Try that. The aim is absolute type compatibility. If types aren't exactly same, the plan is effectively dead. I would say postgresql enforces good habits in it's application developers, from a cultural POV. Had C refused to compile without such strict type compatibility, we wouldn't have to worry about 16bit/32bit and 64 bit software. Just upgrade the compiler and everything is damn good..:-) I doubt if C would have so popular with such strict type checking but that is another issue. I think pascal enforces such strict syntax.. Not sure though.. Shridhar From pgsql-performance-owner@postgresql.org Wed Nov 19 08:24:43 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 7FD49D1D80F for ; Wed, 19 Nov 2003 12:24:37 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 59278-10 for ; Wed, 19 Nov 2003 08:24:08 -0400 (AST) Received: from stan.aopsys.com (m173.net81-64-15.noos.fr [81.64.15.173]) by svr1.postgresql.org (Postfix) with ESMTP id 5999DD1D838 for ; Wed, 19 Nov 2003 08:24:05 -0400 (AST) Received: from localhost ([127.0.0.1] helo=news.nerim.net) by stan.aopsys.com with esmtp (Exim 3.36 #1 (Debian)) id 1AMRFc-0007jT-00; Wed, 19 Nov 2003 13:16:04 +0100 To: Shridhar Daithankar Cc: pgsql-performance@postgresql.org Subject: Re: Join on incompatible types References: <87k75ym1v7.fsf@news.nerim.net> <3FB9F855.7010905@myrealbox.com> <87oev9lsgs.fsf@news.nerim.net> <3FBB138C.4090303@myrealbox.com> <87u150k7mc.fsf@news.nerim.net> <3FBB43AB.30302@myrealbox.com> From: Laurent Martelli Organization: AOPSYS Date: Wed, 19 Nov 2003 13:16:03 +0100 In-Reply-To: <3FBB43AB.30302@myrealbox.com> (Shridhar Daithankar's message of "Wed, 19 Nov 2003 15:49:23 +0530") Message-ID: <87ptfok0zg.fsf@news.nerim.net> User-Agent: Gnus/5.1002 (Gnus v5.10.2) Emacs/20.7 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/155 X-Sequence-Number: 4715 >>>>> "Shridhar" == Shridhar Daithankar writes: [...] Shridhar> 2. Try following query EXPLAIN ANALYZE SELECT * from lists Shridhar> join classes on classes.id=lists.value where Shridhar> lists.id='16'::integer; Shridhar> classes.id=lists.value::integer. With classes.id of type integer and lists.value of type varchar, I get "ERROR: Cannot cast type character varying to integer", which is not such a surprise. Thanks for your help anyway. -- Laurent Martelli laurent@aopsys.com Java Aspect Components http://www.aopsys.com/ http://jac.aopsys.com From pgsql-performance-owner@postgresql.org Wed Nov 19 08:32:28 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id BDBCED1D838 for ; Wed, 19 Nov 2003 12:32:24 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 61503-05 for ; Wed, 19 Nov 2003 08:31:56 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id BA29AD1D83B for ; Wed, 19 Nov 2003 08:31:52 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Wed, 19 Nov 2003 05:31:58 -0700 Message-ID: <3FBB62B5.20507@myrealbox.com> Date: Wed, 19 Nov 2003 18:01:49 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Laurent Martelli Cc: Shridhar Daithankar , pgsql-performance@postgresql.org Subject: Re: Join on incompatible types References: <87k75ym1v7.fsf@news.nerim.net> <3FB9F855.7010905@myrealbox.com> <87oev9lsgs.fsf@news.nerim.net> <3FBB138C.4090303@myrealbox.com> <87u150k7mc.fsf@news.nerim.net> <3FBB43AB.30302@myrealbox.com> <87ptfok0zg.fsf@news.nerim.net> In-Reply-To: <87ptfok0zg.fsf@news.nerim.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/156 X-Sequence-Number: 4716 Laurent Martelli wrote: >>>>>>"Shridhar" == Shridhar Daithankar writes: > > > [...] > > Shridhar> 2. Try following query EXPLAIN ANALYZE SELECT * from lists > Shridhar> join classes on classes.id=lists.value where > Shridhar> lists.id='16'::integer; > > Shridhar> classes.id=lists.value::integer. > > With classes.id of type integer and lists.value of type varchar, I get > "ERROR: Cannot cast type character varying to integer", which is not > such a surprise. Try to_numbr function to get a number out of string. Then cast it to integer. http://developer.postgresql.org/docs/postgres/functions-formatting.html I hope that works. Don't have postgresql installation handy here.. Shridhar From pgsql-performance-owner@postgresql.org Wed Nov 19 10:12:32 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9D3C2D1D861; Wed, 19 Nov 2003 14:11:20 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78780-09; Wed, 19 Nov 2003 10:10:53 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 3B30BD1D859; Wed, 19 Nov 2003 10:10:49 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Wed, 19 Nov 2003 07:10:51 -0700 Message-ID: <3FBB79E6.6040701@myrealbox.com> Date: Wed, 19 Nov 2003 19:40:46 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: josh@agliodbs.com Cc: pgsql-performance@postgresql.org, PostgreSQL-development Subject: Re: More detail on settings for pgavd? References: <200311181558.45847.josh@agliodbs.com> In-Reply-To: <200311181558.45847.josh@agliodbs.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/157 X-Sequence-Number: 4717 Josh Berkus wrote: > Shridhar, > > I was looking at the -V/-v and -A/-a settings in pgavd, and really don't > understand how the calculation works. According to the readme, if I set -v > to 1000 and -V to 2 (the defaults) for a table with 10,000 rows, pgavd would > only vacuum after 21,000 rows had been updated. This seems wrong. No. that is correct. It is calculated as threshold = base + scale*numebr of current rows Which translates to 21,000 = 1000 + 2*1000 However I do not agree with this logic entirely. It pegs the next vacuum w.r.t current table size which is not always a good thing. I would rather vacuum the table at 2000 updates, which is what you probably want. Furthermore analyze threshold depends upon inserts+updates. I think it should also depends upon deletes for obvious reasons. > Can you clear this up a little? I'd like to tweak these settings but can't > without being better aquainted with the calculation. What did you expected in above example? It is not difficult to tweak pg_autovacuum calculations. For testing we can play around. > Also, you may want to reverse your default ratio for Vacuum/analyze frequency. > True, analyze is a less expensive operation than Vacuum, but it's also needed > less often -- only when the *distribution* of data changes. I've seen > databases where the optimal vacuum/analyze frequency was every 10 min/once > per day. OK vacuum and analyze thresholds are calculated with same formula as shown above but with different parameters as follows. vacthresh = vacbase + vacscale*ntuples anathresh = anabase + anascale*ntuples What you are asking for is vacthresh = vacbase*vacscale anathresh = anabase + anascale*ntuples Would that tilt the favour the way you want? i.e. an analyze is triggered when a fixed *percentage* of table changes but a vacuum is triggered when a fixed *number of rows* are changed. I am all for experimentation. If you have real life data to play with, I can give you some patches to play around. And BTW, this is all brain child of Mathew O.Connor(Correct? I am not good at either names or spellings). The way I wrote pgavd originally, each table got to get separate threshold..:-). That was rather a brute force approach. Shridhar From pgsql-performance-owner@postgresql.org Wed Nov 19 12:09:37 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4F7B9D1D58C for ; Wed, 19 Nov 2003 15:51:42 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 00374-05 for ; Wed, 19 Nov 2003 11:51:15 -0400 (AST) Received: from wolff.to (wolff.to [66.93.249.74]) by svr1.postgresql.org (Postfix) with SMTP id A0341D1D587 for ; Wed, 19 Nov 2003 11:51:10 -0400 (AST) Received: (qmail 19165 invoked by uid 500); 19 Nov 2003 15:49:59 -0000 Date: Wed, 19 Nov 2003 09:49:59 -0600 From: Bruno Wolff III To: Rajesh Kumar Mallah Cc: pgsql-performance@postgresql.org Subject: Re: problem with select count(*) .. Message-ID: <20031119154959.GA18327@wolff.to> Mail-Followup-To: Rajesh Kumar Mallah , pgsql-performance@postgresql.org References: <3FBC1ADA.4040709@trade-india.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3FBC1ADA.4040709@trade-india.com> User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/160 X-Sequence-Number: 4720 On Thu, Nov 20, 2003 at 07:07:30 +0530, Rajesh Kumar Mallah wrote: > > If i dump and reload the performance improves and it takes < 1 sec. This > is what i have been doing since the upgrade. But its not a solution. > > The Vacuum full is at the end of a loading batch SQL file which makes lot of > insert , deletes and updates. If a dump and reload fixes your problem, most likely you have a lot of dead tuples in the table. You might need to run vacuum more often. You might have an open transaction that is preventing vacuum full from cleaning up the table. From pgsql-performance-owner@postgresql.org Wed Nov 19 12:07:31 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CC482D1D837 for ; Wed, 19 Nov 2003 16:01:17 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 99676-07 for ; Wed, 19 Nov 2003 12:00:46 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 570C1D1D59B for ; Wed, 19 Nov 2003 12:00:45 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3944212; Wed, 19 Nov 2003 08:01:15 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: Shridhar Daithankar Subject: Re: More detail on settings for pgavd? Date: Wed, 19 Nov 2003 08:00:11 -0800 User-Agent: KMail/1.4.3 Cc: pgsql-performance@postgresql.org References: <200311181558.45847.josh@agliodbs.com> <3FBB0E77.4020105@myrealbox.com> In-Reply-To: <3FBB0E77.4020105@myrealbox.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311190800.11513.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/159 X-Sequence-Number: 4719 Shridhar, > Will look into it. Give me a day or so. I am planning couple of other > patches as well. May be over week end. Thanks, appreciated. As I said, I don't think the settings themselves are wrong, I think the documentation is. What are you patching? -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Wed Nov 19 12:07:25 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 31FA5D1D85C for ; Wed, 19 Nov 2003 16:07:09 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 04794-04 for ; Wed, 19 Nov 2003 12:06:38 -0400 (AST) Received: from mystic.root.hu (mystic.root.hu [62.68.166.4]) by svr1.postgresql.org (Postfix) with ESMTP id 7508BD1D8DE for ; Wed, 19 Nov 2003 12:05:55 -0400 (AST) Received: from [213.163.10.103] (helo=fejleszt4) by mystic.root.hu with asmtp (Exim 3.36 #1 (Debian)) id 1AMUpr-0002ah-00 (authenticated sender: 44b23a02ffc625b094b01917b5736f32) for ; Wed, 19 Nov 2003 17:05:43 +0100 Message-ID: <01e301c3aeb6$bff63350$0403a8c0@fejleszt4> From: "=?iso-8859-2?B?U1rbQ1MgR+Fib3I=?=" To: References: <01a601c3aac1$281e72c0$0403a8c0@fejleszt4> <5187.1068843554@sss.pgh.pa.us> Subject: Re: constant vs function param differs in performance Date: Wed, 19 Nov 2003 17:04:01 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-2" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Authenticated-Sender: 44b23a02ffc625b094b01917b5736f32 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/158 X-Sequence-Number: 4718 Dear Tom, Thanks for your early response. An addition: the nastier difference increased by adding an index (it was an essential index for this query): func with param improved from 2700ms to 2300ms func with constant improved from 400ms to 31ms inline query improved from 390ms to 2ms So am I reading correct and it is completely normal and can't be helped? (couldn't have tried 7.4 yet) In case it reveals something: ------------------------------- cut here ------------------------------- SELECT field FROM (SELECT field, sum(something)=0 AS boolvalue FROM (SELECT * FROM subselect1 NATURAL LEFT JOIN subselect2 UNION SELECT * FROM subselect3 NATURAL LEFT JOIN subselect4 ) AS u GROUP BY field) AS t WHERE not boolvalue ORDER BY simple_sql_func_returns_bool(field) DESC LIMIT 1; ------------------------------- cut here ------------------------------- G. ----- Original Message ----- From: "Tom Lane" Sent: Friday, November 14, 2003 9:59 PM > "=?iso-8859-2?B?U1rbQ1MgR+Fib3I=?=" writes: > > I have two SQL function that produce different times and I can't understand > > why. > > The planner often produces different plans when there are constants in > WHERE clauses than when there are variables, because it can get more > accurate ideas of how many rows will be retrieved. > > regards, tom lane > From pgsql-performance-owner@postgresql.org Wed Nov 19 13:15:19 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 07F96D1B510 for ; Wed, 19 Nov 2003 16:28:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 06681-07 for ; Wed, 19 Nov 2003 12:28:03 -0400 (AST) Received: from email03.aon.at (WARSL402PIP6.highway.telekom.at [195.3.96.93]) by svr1.postgresql.org (Postfix) with SMTP id 6C82CD1D837 for ; Wed, 19 Nov 2003 12:28:02 -0400 (AST) Received: (qmail 562334 invoked from network); 19 Nov 2003 16:27:46 -0000 Received: from m160p013.dipool.highway.telekom.at (HELO cantor) ([62.46.9.237]) (envelope-sender ) by qmail3rs.highway.telekom.at (qmail-ldap-1.03) with SMTP for ; 19 Nov 2003 16:27:46 -0000 From: Manfred Koizar To: Ian Barwick Cc: pgsql-performance@postgresql.org Subject: Re: TEXT column and indexing Date: Wed, 19 Nov 2003 17:26:01 +0100 Message-ID: References: <200311191018.18492.barwick@gmx.net> In-Reply-To: <200311191018.18492.barwick@gmx.net> X-Mailer: Forte Agent 1.93/32.576 English (American) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/162 X-Sequence-Number: 4722 On Wed, 19 Nov 2003 10:18:18 +0100, Ian Barwick wrote: >Indexes: > [...] > "opv_v_ix" btree (substr(value, 1, 128)) >SELECT obj_property_id > FROM object_property_value opv > WHERE opv.value = 'foo' Try ... WHERE substr(opv.value, 1, 128) = 'foo' HTH. Servus Manfred From pgsql-performance-owner@postgresql.org Wed Nov 19 13:12:05 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 91230D1B550 for ; Wed, 19 Nov 2003 16:35:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 09672-06 for ; Wed, 19 Nov 2003 12:35:22 -0400 (AST) Received: from megazone.bigpanda.com (megazone.bigpanda.com [64.147.171.210]) by svr1.postgresql.org (Postfix) with ESMTP id A155DD1D80F for ; Wed, 19 Nov 2003 12:35:21 -0400 (AST) Received: by megazone.bigpanda.com (Postfix, from userid 1001) id 9883A3547C; Wed, 19 Nov 2003 08:35:05 -0800 (PST) Received: from localhost (localhost [127.0.0.1]) by megazone.bigpanda.com (Postfix) with ESMTP id 96FC935466; Wed, 19 Nov 2003 08:35:05 -0800 (PST) Date: Wed, 19 Nov 2003 08:35:05 -0800 (PST) From: Stephan Szabo To: Ian Barwick Cc: pgsql-performance@postgresql.org Subject: Re: TEXT column and indexing In-Reply-To: <200311191018.18492.barwick@gmx.net> Message-ID: <20031119083025.R85482@megazone.bigpanda.com> References: <200311191018.18492.barwick@gmx.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/161 X-Sequence-Number: 4721 On Wed, 19 Nov 2003, Ian Barwick wrote: > > I have this table: > > db=> \d object_property_value > Table "db.object_property_value" > Column | Type | Modifiers > -----------------------+------------------------+-------------------- > obj_property_value_id | integer | not null default nextval(... > obj_property_id | integer | not null > value | text | > Indexes: > "object_property_value_pkey" primary key, btree (obj_property_value_id) > "opv_obj_property_id_ix" btree (obj_property_id) > "opv_v_ix" btree (substr(value, 1, 128)) > Foreign-key constraints: > "object_property_fkey" FOREIGN KEY (obj_property_id) > REFERENCES object_property(obj_property_id) > ON UPDATE CASCADE ON DELETE CASCADE > I want to query this table to match a specific value along > the lines of: > > SELECT obj_property_id > FROM object_property_value opv > WHERE opv.value = 'foo' > > The question is therefore: can I get an index to work on the TEXT column? It > is currently indexed with: > "opv_v_ix" btree (substr(value, 1, 128)) > > which doesn't appear to have any effect. I am probably missing something > obvious though. I can live with maintaining an extra VARCHAR column but You probably need to be querying like: WHERE substr(value,1,128)='foo'; in order to use that index. While substr(txtcol, 1,128) happens to have the property that it would be probably be useful in a search against a short constant string, that's an internal property of that function. From pgsql-performance-owner@postgresql.org Wed Nov 19 13:28:40 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C1416D1B50F; Wed, 19 Nov 2003 17:08:28 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 22433-01; Wed, 19 Nov 2003 13:07:58 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 75B53D1CB08; Wed, 19 Nov 2003 13:06:33 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3944461; Wed, 19 Nov 2003 09:07:20 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: Shridhar Daithankar Subject: Re: More detail on settings for pgavd? Date: Wed, 19 Nov 2003 09:06:15 -0800 User-Agent: KMail/1.4.3 Cc: pgsql-performance@postgresql.org, PostgreSQL-development References: <200311181558.45847.josh@agliodbs.com> <3FBB79E6.6040701@myrealbox.com> In-Reply-To: <3FBB79E6.6040701@myrealbox.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311190906.15828.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/163 X-Sequence-Number: 4723 Shridhar, > However I do not agree with this logic entirely. It pegs the next vacuum > w.r.t current table size which is not always a good thing. No, I think the logic's fine, it's the numbers which are wrong. We want to vacuum when updates reach between 5% and 15% of total rows. NOT when updates reach 110% of total rows ... that's much too late. Hmmm ... I also think the threshold level needs to be lowered; I guess the purpose was to prevent continuous re-vacuuuming of small tables? Unfortunately, in the current implementation, the result is tha small tables never get vacuumed at all. So for defaults, I would peg -V at 0.1 and -v at 100, so our default calculation for a table with 10,000 rows is: 100 + ( 0.1 * 10,000 ) = 1100 rows. > I would rather vacuum the table at 2000 updates, which is what you probably > want. Not necessarily. This would be painful if the table has 10,000,000 rows. It *should* be based on a % of rows. > Furthermore analyze threshold depends upon inserts+updates. I think it > should also depends upon deletes for obvious reasons. Yes. Vacuum threshold is counting deletes, I hope? > What did you expected in above example? It is not difficult to tweak > pg_autovacuum calculations. For testing we can play around. Can I set the settings to decimals, or are they integers? > vacthresh = vacbase*vacscale > anathresh = anabase + anascale*ntuples Nope, see above. My comment about the frequency of vacuums vs. analyze is that currently the *default* is to analyze twice as often as you vacuum. Based on my experiece as a PG admin on a variety of databases, I believe that the default should be to analyze half as often as you vacuum. > I am all for experimentation. If you have real life data to play with, I > can give you some patches to play around. I will have real data very soon ..... -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Wed Nov 19 15:26:29 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 020C5D1C50F for ; Wed, 19 Nov 2003 18:39:18 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 35921-04 for ; Wed, 19 Nov 2003 14:38:47 -0400 (AST) Received: from nitrogen.id.pl (nitrogen.id.pl [193.178.214.5]) by svr1.postgresql.org (Postfix) with SMTP id 218FED1B578 for ; Wed, 19 Nov 2003 14:38:44 -0400 (AST) Received: (qmail 26079 invoked by uid 0); 19 Nov 2003 18:38:28 -0000 Received: from unknown (HELO siaco.id.pl) (213.25.114.8) by smtp.id.pl with SMTP; 19 Nov 2003 18:38:28 -0000 Received: (qmail 19544 invoked by uid 1000); 19 Nov 2003 18:38:24 -0000 Date: Wed, 19 Nov 2003 19:38:24 +0100 From: Ryszard Lach To: Bruce Momjian Cc: pgsql-performance@postgresql.org Subject: Re: duration logging setting in 7.4 Message-ID: <20031119183824.GL18018@siaco.id.pl> Reply-To: Ryszard Lach References: <20031118091646.GA8223@siaco.id.pl> <200311181507.hAIF7mX04577@candle.pha.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-2 Content-Disposition: inline In-Reply-To: <200311181507.hAIF7mX04577@candle.pha.pa.us> X-My-GPG-Key: echo | mail -s "send key pub" ryszard@lach.name User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/167 X-Sequence-Number: 4727 On Tue, Nov 18, 2003 at 10:07:48AM -0500, Bruce Momjian wrote: > > Wow, that is strange. If you don't use syslog, do you see the proper > output? I've just checked this. It behaves exactly the same way. > If you turn on log_statement, do you see the statements? If I turn on log_min_duration_statement (i.e. set to 0), log_statement and log_duration, then I receive something like that Nov 17 22:33:27 postgres[22945]: [29231-1] LOG: statement: Nov 17 22:33:27 postgres[22945]: [29232-1] LOG: duration: 0.198 ms Nov 17 22:33:27 postgres[22945]: [29233-1] LOG: duration: 0.198 ms statement: Nov 17 22:33:27 postgres[22946]: [29231-1] LOG: statement: Nov 17 22:33:27 postgres[22946]: [29232-1] LOG: duration: 0.191 ms Nov 17 22:33:27 postgres[22946]: [29233-1] LOG: duration: 0.191 ms statement: Nov 17 22:33:27 postgres[22678]: [147134-1] LOG: statement: select * from cms where id=1465 Nov 17 22:33:27 postgres[22679]: [154907-1] LOG: statement: Nov 17 22:33:27 postgres[22679]: [154908-1] LOG: duration: 0.867 ms Nov 17 22:33:27 postgres[22679]: [154909-1] LOG: duration: 0.867 ms statement: Nov 17 22:33:27 postgres[22678]: [147135-1] LOG: duration: 1.458 ms Nov 17 22:33:27 postgres[22678]: [147136-1] LOG: duration: 1.458 ms statement: select * from cms where id=1465 Nov 17 22:33:27 postgres[22680]: [158366-1] LOG: statement: Nov 17 22:33:27 postgres[22680]: [158367-1] LOG: duration: 0.620 ms Nov 17 22:33:27 postgres[22680]: [158368-1] LOG: duration: 0.620 ms statement: Nov 17 22:33:27 postgres[22681]: [161294-1] LOG: statement: Nov 17 22:33:27 postgres[22681]: [161295-1] LOG: duration: 0.650 ms It seems, that log_duration is responsible only for "duration:" lines, log_statement - for "statement:" ones, and "log_min_duration_statement" - for "duration: .* statement:". I think, that the above output should exclude losing of data by syslog from further delibarations. Do you thing that could be a bug? There is another one thing: logs from the same database running on 7.3 and the same application contained lines like 'select getdatabaseencoding()', 'select datestyle()' and similar (not used by application explicite, probably added by JDBC driver), now they are missed - maybe this is the problem? Richard. -- "First they ignore you. Then they laugh at you. Then they fight you. Then you win." - Mohandas Gandhi. From pgsql-performance-owner@postgresql.org Wed Nov 19 14:59:04 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EBE4DD1B57C for ; Wed, 19 Nov 2003 18:59:02 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38736-06 for ; Wed, 19 Nov 2003 14:58:32 -0400 (AST) Received: from candle.pha.pa.us (unknown [207.106.42.251]) by svr1.postgresql.org (Postfix) with ESMTP id 716DFD1BB42 for ; Wed, 19 Nov 2003 14:58:29 -0400 (AST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.11.6) id hAJIwR810214; Wed, 19 Nov 2003 13:58:27 -0500 (EST) From: Bruce Momjian Message-Id: <200311191858.hAJIwR810214@candle.pha.pa.us> Subject: Re: duration logging setting in 7.4 In-Reply-To: <20031119183824.GL18018@siaco.id.pl> To: Ryszard Lach Date: Wed, 19 Nov 2003 13:58:27 -0500 (EST) Cc: pgsql-performance@postgresql.org X-Mailer: ELM [version 2.4ME+ PL108 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/164 X-Sequence-Number: 4724 Ryszard Lach wrote: > If I turn on log_min_duration_statement (i.e. set to 0), log_statement and > log_duration, then I receive something like that > > Nov 17 22:33:27 postgres[22945]: [29231-1] LOG: statement: > Nov 17 22:33:27 postgres[22945]: [29232-1] LOG: duration: 0.198 ms > Nov 17 22:33:27 postgres[22945]: [29233-1] LOG: duration: 0.198 ms statement: > Nov 17 22:33:27 postgres[22946]: [29231-1] LOG: statement: > Nov 17 22:33:27 postgres[22946]: [29232-1] LOG: duration: 0.191 ms > Nov 17 22:33:27 postgres[22946]: [29233-1] LOG: duration: 0.191 ms statement: > Nov 17 22:33:27 postgres[22678]: [147134-1] LOG: statement: select * from cms where id=1465 > Nov 17 22:33:27 postgres[22679]: [154907-1] LOG: statement: > Nov 17 22:33:27 postgres[22679]: [154908-1] LOG: duration: 0.867 ms > Nov 17 22:33:27 postgres[22679]: [154909-1] LOG: duration: 0.867 ms statement: > Nov 17 22:33:27 postgres[22678]: [147135-1] LOG: duration: 1.458 ms > Nov 17 22:33:27 postgres[22678]: [147136-1] LOG: duration: 1.458 ms statement: select * from cms where id=1465 > Nov 17 22:33:27 postgres[22680]: [158366-1] LOG: statement: > Nov 17 22:33:27 postgres[22680]: [158367-1] LOG: duration: 0.620 ms > Nov 17 22:33:27 postgres[22680]: [158368-1] LOG: duration: 0.620 ms statement: > Nov 17 22:33:27 postgres[22681]: [161294-1] LOG: statement: > Nov 17 22:33:27 postgres[22681]: [161295-1] LOG: duration: 0.650 ms > > It seems, that log_duration is responsible only for "duration:" lines, > log_statement - for "statement:" ones, and "log_min_duration_statement" - for > "duration: .* statement:". I think, that the above output should exclude losing > of data by syslog from further delibarations. Do you thing that could be > a bug? Yes, the problem is not related to syslog. Are you using prepared queries, perhaps? I don't think those show the query, but it seems we should display something better than blanks. > There is another one thing: logs from the same database running on 7.3 and the same > application contained lines like 'select getdatabaseencoding()', 'select > datestyle()' and similar (not used by application explicite, probably > added by JDBC driver), now they are missed - maybe this is the > problem? No, those are missing because the new 7.4 wire protocol doesn't require those queries anymore --- the data is send automatically. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073 From pgsql-performance-owner@postgresql.org Wed Nov 19 15:15:30 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D5FC6D1CCAC for ; Wed, 19 Nov 2003 19:15:23 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45719-03 for ; Wed, 19 Nov 2003 15:14:53 -0400 (AST) Received: from p15140235.pureserver.info (sql-info.de [217.160.166.164]) by svr1.postgresql.org (Postfix) with ESMTP id 11F0DD1BB13 for ; Wed, 19 Nov 2003 15:14:52 -0400 (AST) Received: by p15140235.pureserver.info (Postfix, from userid 639) id E38EA203A; Wed, 19 Nov 2003 20:14:50 +0100 (CET) Received: from iwate.sql-info.de (pD9517DFE.dip.t-dialin.net [217.81.125.254]) (using SSLv3 with cipher RC4-MD5 (128/128 bits)) (No client certificate requested) by p15140235.pureserver.info (Postfix) with ESMTP id 03C1CEA1; Wed, 19 Nov 2003 20:14:49 +0100 (CET) From: Ian Barwick To: Stephan Szabo Subject: Re: TEXT column and indexing Date: Wed, 19 Nov 2003 20:13:42 +0100 User-Agent: KMail/1.5.1 Cc: pgsql-performance@postgresql.org References: <200311191018.18492.barwick@gmx.net> <20031119083025.R85482@megazone.bigpanda.com> In-Reply-To: <20031119083025.R85482@megazone.bigpanda.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311192013.42574.barwick@gmx.net> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/165 X-Sequence-Number: 4725 On Wednesday 19 November 2003 17:35, Stephan Szabo wrote: > On Wed, 19 Nov 2003, Ian Barwick wrote: > > I have this table: (...) > > You probably need to be querying like: > WHERE substr(value,1,128)='foo'; > in order to use that index. > > While substr(txtcol, 1,128) happens to have the property that it would be > probably be useful in a search against a short constant string, that's an > internal property of that function. That's the one :-). Thanks! Ian Barwick barwick@gmx.net From pgsql-performance-owner@postgresql.org Wed Nov 19 15:17:24 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E595DD1CCDE for ; Wed, 19 Nov 2003 19:17:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 38718-08 for ; Wed, 19 Nov 2003 15:16:52 -0400 (AST) Received: from p15140235.pureserver.info (sql-info.de [217.160.166.164]) by svr1.postgresql.org (Postfix) with ESMTP id 16D3DD1B53A for ; Wed, 19 Nov 2003 15:16:51 -0400 (AST) Received: by p15140235.pureserver.info (Postfix, from userid 639) id DA3BF203A; Wed, 19 Nov 2003 20:16:51 +0100 (CET) Received: from iwate.sql-info.de (pD9517DFE.dip.t-dialin.net [217.81.125.254]) (using SSLv3 with cipher RC4-MD5 (128/128 bits)) (No client certificate requested) by p15140235.pureserver.info (Postfix) with ESMTP id AD4EEEA1; Wed, 19 Nov 2003 20:16:50 +0100 (CET) From: Ian Barwick To: Manfred Koizar Subject: Re: TEXT column and indexing Date: Wed, 19 Nov 2003 20:16:48 +0100 User-Agent: KMail/1.5.1 References: <200311191018.18492.barwick@gmx.net> In-Reply-To: Cc: pgsql-performance@postgresql.org MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311192016.48447.barwick@gmx.net> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/166 X-Sequence-Number: 4726 On Wednesday 19 November 2003 17:26, you wrote: > On Wed, 19 Nov 2003 10:18:18 +0100, Ian Barwick > > wrote: > >Indexes: > > [...] > > "opv_v_ix" btree (substr(value, 1, 128)) > > > >SELECT obj_property_id > > FROM object_property_value opv > > WHERE opv.value = 'foo' > > Try > ... WHERE substr(opv.value, 1, 128) = 'foo' > > HTH. Yup: db=> explain db-> SELECT obj_property_id db-> FROM object_property_value opv db-> WHERE substr(opv.value,1,128) = 'foo'; QUERY PLAN ------------------------------------------------------------------------------------------------ Index Scan using opv_v_ix on object_property_value opv (cost=0.00..4185.78 rows=1101 width=4) Index Cond: (substr(value, 1, 128) = 'foo'::text) (2 rows) Many thanks Ian Barwick barwick@gmx.net From pgsql-performance-owner@postgresql.org Wed Nov 19 15:34:17 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 412B5D1B523 for ; Wed, 19 Nov 2003 19:29:32 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 47470-02 for ; Wed, 19 Nov 2003 15:29:01 -0400 (AST) Received: from nitrogen.id.pl (nitrogen.id.pl [193.178.214.5]) by svr1.postgresql.org (Postfix) with SMTP id 53390D1C96A for ; Wed, 19 Nov 2003 15:28:59 -0400 (AST) Received: (qmail 8593 invoked by uid 0); 19 Nov 2003 19:28:59 -0000 Received: from unknown (HELO siaco.id.pl) (213.25.114.8) by smtp.id.pl with SMTP; 19 Nov 2003 19:28:59 -0000 Received: (qmail 19907 invoked by uid 1000); 19 Nov 2003 19:28:55 -0000 Date: Wed, 19 Nov 2003 20:28:55 +0100 From: Ryszard Lach To: Bruce Momjian Cc: pgsql-performance@postgresql.org Subject: Re: duration logging setting in 7.4 Message-ID: <20031119192855.GO18018@siaco.id.pl> Reply-To: Ryszard Lach References: <20031119183824.GL18018@siaco.id.pl> <200311191858.hAJIwR810214@candle.pha.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-2 Content-Disposition: inline In-Reply-To: <200311191858.hAJIwR810214@candle.pha.pa.us> X-My-GPG-Key: echo | mail -s "send key pub" ryszard@lach.name User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/168 X-Sequence-Number: 4728 On Wed, Nov 19, 2003 at 01:58:27PM -0500, Bruce Momjian wrote: > Ryszard Lach wrote: > > > There is another one thing: logs from the same database running on 7.3 and the same > > application contained lines like 'select getdatabaseencoding()', 'select > > datestyle()' and similar (not used by application explicite, probably > > added by JDBC driver), now they are missed - maybe this is the > > problem? > > No, those are missing because the new 7.4 wire protocol doesn't require > those queries anymore --- the data is send automatically. > Mayby this is a solution? Because of some charset-related problems we are still using an old (AFAiR modified) version of JDBC driver. I'm not a programmer, but I think and don't know what JDBC driver does, but maybe it sends from client side those queries and server doesn't know what to do with them? I'll ask our programmers to try with 7.4 driver and tell you about results. Richard. -- "First they ignore you. Then they laugh at you. Then they fight you. Then you win." - Mohandas Gandhi. From pgsql-performance-owner@postgresql.org Wed Nov 19 15:45:28 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 22E74D1C96A for ; Wed, 19 Nov 2003 19:42:58 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 48862-05 for ; Wed, 19 Nov 2003 15:42:27 -0400 (AST) Received: from candle.pha.pa.us (unknown [207.106.42.251]) by svr1.postgresql.org (Postfix) with ESMTP id 454BFD1D279 for ; Wed, 19 Nov 2003 15:42:25 -0400 (AST) Received: (from pgman@localhost) by candle.pha.pa.us (8.11.6/8.11.6) id hAJJgKR15032; Wed, 19 Nov 2003 14:42:20 -0500 (EST) From: Bruce Momjian Message-Id: <200311191942.hAJJgKR15032@candle.pha.pa.us> Subject: Re: duration logging setting in 7.4 In-Reply-To: <20031119192855.GO18018@siaco.id.pl> To: Ryszard Lach Date: Wed, 19 Nov 2003 14:42:20 -0500 (EST) Cc: pgsql-performance@postgresql.org X-Mailer: ELM [version 2.4ME+ PL108 (25)] MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/169 X-Sequence-Number: 4729 Ryszard Lach wrote: > On Wed, Nov 19, 2003 at 01:58:27PM -0500, Bruce Momjian wrote: > > Ryszard Lach wrote: > > > > > There is another one thing: logs from the same database running on 7.3 and the same > > > application contained lines like 'select getdatabaseencoding()', 'select > > > datestyle()' and similar (not used by application explicite, probably > > > added by JDBC driver), now they are missed - maybe this is the > > > problem? > > > > No, those are missing because the new 7.4 wire protocol doesn't require > > those queries anymore --- the data is send automatically. > > > > Mayby this is a solution? Because of some > charset-related problems we are still using an old (AFAiR modified) > version of JDBC driver. I'm not a programmer, but I think and don't know > what JDBC driver does, but maybe it sends from client side those queries > and server doesn't know what to do with them? I'll ask our programmers > to try with 7.4 driver and tell you about results. Also, try plain psql and issue a query and see if it appears. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073 From pgsql-performance-owner@postgresql.org Tue Nov 18 21:37:51 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A4D84D1D84D for ; Wed, 19 Nov 2003 01:37:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 60988-05 for ; Tue, 18 Nov 2003 21:37:12 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id A6370D1D630 for ; Tue, 18 Nov 2003 21:36:58 -0400 (AST) Received: (qmail 18951 invoked from network); 19 Nov 2003 01:38:05 -0000 Received: from unknown (HELO trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 19 Nov 2003 01:38:05 -0000 Message-ID: <3FBC1ADA.4040709@trade-india.com> Date: Thu, 20 Nov 2003 07:07:30 +0530 From: Rajesh Kumar Mallah User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030630 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: problem with select count(*) .. Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/148 X-Sequence-Number: 4708 Ever Since i upgraded to 7.4RC2 i am facing problem with select count(*) . In 7.3 the problem was not there select count(*) from data_bank.profiles used to return almost instantly , but in 7.4 explain analyze SELECT count(*) from data_bank.profiles; QUERY PLAN --------------------------------------------------------------------------- Aggregate (cost=48361.30..48361.30 rows=1 width=0) (actual time=23456.870..23456.871 rows=1 loops=1) -> Seq Scan on profiles (cost=0.00..47431.84 rows=371784 width=0) (actual time=12174.999..23262.823 rows=123928 loops=1) Total runtime: 23458.460 ms (3 rows) tradein_clients=# If i dump and reload the performance improves and it takes < 1 sec. This is what i have been doing since the upgrade. But its not a solution. The Vacuum full is at the end of a loading batch SQL file which makes lot of insert , deletes and updates. Regds Mallah. VACUUM FULL VERBOSE ANALYZE data_bank.profiles; INFO: vacuuming "data_bank.profiles" INFO: "profiles": found 430524 removable, 371784 nonremovable row versions in 43714 pages INFO: index "profiles_pincode" now contains 371784 row versions in 3419 pages INFO: index "profiles_city" now contains 371784 row versions in 3471 pages INFO: index "profiles_branch" now contains 371784 row versions in 2237 pages INFO: index "profiles_area_code" now contains 371784 row versions in 2611 pages INFO: index "profiles_source" now contains 371784 row versions in 3165 pages INFO: index "co_name_index_idx" now contains 371325 row versions in 3933 pages INFO: index "address_index_idx" now contains 371490 row versions in 4883 pages INFO: index "profiles_exp_cat" now contains 154836 row versions in 2181 pages INFO: index "profiles_imp_cat" now contains 73678 row versions in 1043 pages INFO: index "profiles_manu_cat" now contains 87124 row versions in 1201 pages INFO: index "profiles_serv_cat" now contains 19340 row versions in 269 pages INFO: index "profiles_pid" now contains 371784 row versions in 817 pages INFO: index "profiles_pending_branch_id" now contains 0 row versions in 1 pages INFO: "profiles": moved 0 row versions, truncated 43714 to 43714 pages INFO: vacuuming "pg_toast.pg_toast_67748379" INFO: "pg_toast_67748379": found 0 removable, 74 nonremovable row versions in 17 pages INFO: index "pg_toast_67748379_index" now contains 74 row versions in 2 pages INFO: "pg_toast_67748379": moved 1 row versions, truncated 17 to 17 pages INFO: index "pg_toast_67748379_index" now contains 74 row versions in 2 pages INFO: analyzing "data_bank.profiles" INFO: "profiles": 43714 pages, 3000 rows sampled, 3634 estimated total rows VACUUM Time: 1001525.19 ms From pgsql-performance-owner@postgresql.org Thu Nov 20 00:40:32 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DD402D1B506 for ; Thu, 20 Nov 2003 04:40:25 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 37911-08 for ; Thu, 20 Nov 2003 00:39:55 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 5B510D1B4F6 for ; Thu, 20 Nov 2003 00:39:54 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAK4dl19001172; Wed, 19 Nov 2003 23:39:47 -0500 (EST) To: Joel Jacobson Cc: pgsql-performance@postgresql.org Subject: Re: Backup/restore of pg_statistics In-reply-to: <1069080680.3fb8e068eb941@mail.jacobson.be> References: <1069080680.3fb8e068eb941@mail.jacobson.be> Comments: In-reply-to Joel Jacobson message dated "Mon, 17 Nov 2003 15:51:20 +0100" Date: Wed, 19 Nov 2003 23:39:47 -0500 Message-ID: <1171.1069303187@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/170 X-Sequence-Number: 4730 Joel Jacobson writes: > I understand that it is not possible to occasionally re-plan the queries in a > PL/pgSQL function without dropping and re-creating the function. Huh? You only need to start a fresh connection. regards, tom lane From pgsql-performance-owner@postgresql.org Thu Nov 20 01:58:52 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 20B42D1CA63 for ; Thu, 20 Nov 2003 05:58:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 59838-01 for ; Thu, 20 Nov 2003 01:58:20 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id E1F9DD1B592 for ; Thu, 20 Nov 2003 01:58:18 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Wed, 19 Nov 2003 22:58:19 -0700 Message-ID: <3FBC57F5.8070705@myrealbox.com> Date: Thu, 20 Nov 2003 11:28:13 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Benjamin Bostow , Performance Subject: Re: High Processor consumption References: <1069204784.1955.16.camel@localhost.localdomain> <3FBB0EE4.6010805@myrealbox.com> <1069264247.1955.25.camel@localhost.localdomain> In-Reply-To: <1069264247.1955.25.camel@localhost.localdomain> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/171 X-Sequence-Number: 4731 Benjamin Bostow wrote: > I haven't modified any of the setting. I did try changing shmmax from > 32MB to 256MB but didn't see much change in the processor usage. The > init script that runs to start the server uses the following: > su -l postgres -s /bin/sh -c "/usr/bin/pg_ctl -D $PGDATA -p > /usr/bin/postmaster -o \"-i -N 128 -B 256\" start > /dev/null 2>&1" < > /dev/null Lol.. -B 256 does not mean 256MB of buffers. It means 2MB of buffers. Each buffer is of 8K. Try upping it to a 1000 or 2000 > > I haven't modified postgresql.conf yet but am going through a book on > performance tuning the server. If you can provide any suggestions or > help as I am new to postgres it would be appreciated. http://www.varlena.com/varlena/GeneralBits/Tidbits/perf.html HTH Shridhar From pgsql-performance-owner@postgresql.org Thu Nov 20 03:24:11 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 520C1D1C9C0; Thu, 20 Nov 2003 07:24:06 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 76556-01; Thu, 20 Nov 2003 03:23:36 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 75C9DD1C9E5; Thu, 20 Nov 2003 03:23:33 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Thu, 20 Nov 2003 00:23:33 -0700 Message-ID: <3FBC6BED.9090809@myrealbox.com> Date: Thu, 20 Nov 2003 12:53:25 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Josh Berkus Cc: pgsql-performance@postgresql.org, PostgreSQL-development Subject: Re: More detail on settings for pgavd? References: <200311181558.45847.josh@agliodbs.com> <3FBB79E6.6040701@myrealbox.com> <200311190906.15828.josh@agliodbs.com> In-Reply-To: <200311190906.15828.josh@agliodbs.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/172 X-Sequence-Number: 4732 Josh Berkus wrote: > Shridhar, >>However I do not agree with this logic entirely. It pegs the next vacuum >>w.r.t current table size which is not always a good thing. > > > No, I think the logic's fine, it's the numbers which are wrong. We want to > vacuum when updates reach between 5% and 15% of total rows. NOT when > updates reach 110% of total rows ... that's much too late. Well, looks like thresholds below 1 should be norm rather than exception. > Hmmm ... I also think the threshold level needs to be lowered; I guess the > purpose was to prevent continuous re-vacuuuming of small tables? > Unfortunately, in the current implementation, the result is tha small tables > never get vacuumed at all. > > So for defaults, I would peg -V at 0.1 and -v at 100, so our default > calculation for a table with 10,000 rows is: > > 100 + ( 0.1 * 10,000 ) = 1100 rows. I would say -V 0.2-0.4 could be great as well. Fact to emphasize is that thresholds less than 1 should be used. >>Furthermore analyze threshold depends upon inserts+updates. I think it >>should also depends upon deletes for obvious reasons. > Yes. Vacuum threshold is counting deletes, I hope? It does. > My comment about the frequency of vacuums vs. analyze is that currently the > *default* is to analyze twice as often as you vacuum. Based on my > experiece as a PG admin on a variety of databases, I believe that the default > should be to analyze half as often as you vacuum. OK. >>I am all for experimentation. If you have real life data to play with, I >>can give you some patches to play around. > I will have real data very soon ..... I will submit a patch that would account deletes in analyze threshold. Since you want to delay the analyze, I would calculate analyze count as n=updates + inserts *-* deletes Rather than current "n = updates + inserts". Also update readme about examples and analyze frequency. What does statistics gather BTW? Just number of rows or something else as well? I think I would put that on Hackers separately. I am still wary of inverting vacuum analyze frequency. You think it is better to set inverted default rather than documenting it? Shridhar From pgsql-hackers-owner@postgresql.org Thu Nov 20 10:31:58 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3A209D1B506; Thu, 20 Nov 2003 14:31:50 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 42543-02; Thu, 20 Nov 2003 10:31:22 -0400 (AST) Received: from mta7.srv.hcvlny.cv.net (mta7.srv.hcvlny.cv.net [167.206.5.74]) by svr1.postgresql.org (Postfix) with ESMTP id 870C2D1B520; Thu, 20 Nov 2003 10:31:12 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta7.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HON00ERTMZ3G2@mta7.srv.hcvlny.cv.net>; Thu, 20 Nov 2003 09:30:42 -0500 (EST) Date: Thu, 20 Nov 2003 09:30:43 -0500 From: "Matthew T. O'Connor" Subject: Re: [PERFORM] More detail on settings for pgavd? In-reply-to: <3FBC6BED.9090809@myrealbox.com> To: Shridhar Daithankar Cc: Josh Berkus , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBCD013.3030402@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <200311181558.45847.josh@agliodbs.com> <3FBB79E6.6040701@myrealbox.com> <200311190906.15828.josh@agliodbs.com> <3FBC6BED.9090809@myrealbox.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/1117 X-Sequence-Number: 47405 Shridhar Daithankar wrote: > Josh Berkus wrote: > >> Shridhar, > > >>However I do not agree with this logic entirely. It pegs the next > vacuum > >>> w.r.t current table size which is not always a good thing. >> Ok, what do you recommend? The point of two separate variables allows you to specify if you want vacuum based on a fixed number, based on table size or something inbetween. >> >> No, I think the logic's fine, it's the numbers which are wrong. We >> want to vacuum when updates reach between 5% and 15% of total rows. >> NOT when updates reach 110% of total rows ... that's much too late. > For small tables, you don't need to vacuum too often. In the testing I did a small table ~100 rows, didn't really show significant performance degredation until it had close to 1000 updates. For large tables, vacuum is so expensive, that you don't want to do it very often, and scanning the whole table when there is only 5% wasted space is not very helpful. >> Hmmm ... I also think the threshold level needs to be lowered; I >> guess the purpose was to prevent continuous re-vacuuuming of small >> tables? Unfortunately, in the current implementation, the result is >> tha small tables never get vacuumed at all. >> >> So for defaults, I would peg -V at 0.1 and -v at 100, so our default >> calculation for a table with 10,000 rows is: >> >> 100 + ( 0.1 * 10,000 ) = 1100 rows. > Yes, the I set the defaults a little high perhaps so as to err on the side of caution. I didn't want people to say pg_autovacuum kills the performance of my server. A small table will get vacuumed, just not until it has reached the threshold. So a table with 100 rows, will get vacuumed after 1200 updates / deletes. In my testing it showed that there was no major performance problems until you reached several thousand updates / deletes. >>> Furthermore analyze threshold depends upon inserts+updates. I think it >>> should also depends upon deletes for obvious reasons. >> >> Yes. Vacuum threshold is counting deletes, I hope? > > It does. > >> My comment about the frequency of vacuums vs. analyze is that >> currently the *default* is to analyze twice as often as you >> vacuum. Based on my experiece as a PG admin on a variety of >> databases, I believe that the default should be to analyze half as >> often as you vacuum. > HUH? analyze is very very cheap compared to vacuum. Why not do it more often? >>> I am all for experimentation. If you have real life data to play >>> with, I >>> can give you some patches to play around. >> >> I will have real data very soon ..... > > I will submit a patch that would account deletes in analyze threshold. > Since you want to delay the analyze, I would calculate analyze count as deletes are already accounted for in the analyze threshold. > I am still wary of inverting vacuum analyze frequency. You think it is > better to set inverted default rather than documenting it? I think inverting the vacuum and analyze frequency is wrong. What I think I am hearing is that people would like very much to be able to tweak the settings of pg_autovacuum for individual tables / databases etc. So that you could set certain tables to be vacuumed more agressivly than others. I agree this would be a good and welcome addition. I hope have time to work on this at some point, but in the near future I won't. Matthew From pgsql-performance-owner@postgresql.org Thu Nov 20 11:00:59 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 805E0D1B4E6; Thu, 20 Nov 2003 15:00:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 42189-09; Thu, 20 Nov 2003 11:00:24 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 99A8BD1B50A; Thu, 20 Nov 2003 11:00:20 -0400 (AST) Received: from daithan.intranet.pspl.co.in shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Thu, 20 Nov 2003 08:00:08 -0700 From: Shridhar Daithankar To: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? Date: Thu, 20 Nov 2003 20:29:47 +0530 User-Agent: KMail/1.5.4 Cc: Josh Berkus , pgsql-performance@postgresql.org, PostgreSQL-development References: <200311181558.45847.josh@agliodbs.com> <3FBC6BED.9090809@myrealbox.com> <3FBCD013.3030402@zeut.net> In-Reply-To: <3FBCD013.3030402@zeut.net> MIME-Version: 1.0 Content-Type: Multipart/Mixed; boundary="Boundary-00=_jbNv/4RC6mqXtJy" Message-Id: <200311202029.47351.shridhar_daithankar@myrealbox.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/173 X-Sequence-Number: 4733 --Boundary-00=_jbNv/4RC6mqXtJy Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline On Thursday 20 November 2003 20:00, Matthew T. O'Connor wrote: > Shridhar Daithankar wrote: > > I will submit a patch that would account deletes in analyze threshold. > > Since you want to delay the analyze, I would calculate analyze count as > > deletes are already accounted for in the analyze threshold. Yes. My bad. Deletes are not accounted in initializing analyze count but later they are used. > > I am still wary of inverting vacuum analyze frequency. You think it is > > better to set inverted default rather than documenting it? > > I think inverting the vacuum and analyze frequency is wrong. Me. Too. ATM all I can think of this patch attached. Josh, is it sufficient for you?..:-) Matthew, I am confyused about one thing. Why would autovacuum count updates while checking for analyze threshold? Analyze does not change statistics right? ( w.r.t line 1072, pg_autovacuum.c). For updating statistics, only inserts+deletes should suffice, isn't it? Other than that, I think autovacuum does everything it can. Comments? Shridhar --Boundary-00=_jbNv/4RC6mqXtJy Content-Type: text/x-diff; charset="iso-8859-1"; name="difffile" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="difffile" *** README.pg_autovacuum.orig Thu Nov 20 19:58:29 2003 --- README.pg_autovacuum Thu Nov 20 20:26:39 2003 *************** *** 141,150 **** depending on the mixture of table activity (insert, update, or delete): ! - If the number of (inserts + updates + deletes) > AnalyzeThreshold, then only an analyze is performed. ! - If the number of (deletes + updates) > VacuumThreshold, then a vacuum analyze is performed. VacuumThreshold is equal to: --- 141,150 ---- depending on the mixture of table activity (insert, update, or delete): ! - If the number of (inserts + updates + deletes) >= AnalyzeThreshold, then only an analyze is performed. ! - If the number of (deletes + updates) >= VacuumThreshold, then a vacuum analyze is performed. VacuumThreshold is equal to: *************** *** 158,163 **** --- 158,186 ---- and running ANALYZE more often should not substantially degrade system performance. + Examples: + + Following table shows typical usage of pg_autovacuum settings. + These are put here so that a DBA can have some starting point while + tuning pg_autovacuum. + + Vacuum is triggered by updates and deletes. So in case of vacuum, + last column indicates total of updates and deletes required + to trigger vacuum. In case of analyze, the operations would count total + number of inserts, updates and deletes. + + Threshold Scaling factor Records No. of Operations + 1,000 1 10,000 11,000 + 1,000 2 10,000 21,000 + 1,000 0.5 10,000 6,000 + 1,000 0.1 10,000 2,000 + + Although analyze is cheaper operation compared to vacuum, + it might be needed less often. The default is to analyze twice as much as + vacuum but that might be too aggressive for some installations. It is advised that + such installation tune their analyze threshold separately, rather than relying upon + the default behaviour. + Sleeping: --------- --Boundary-00=_jbNv/4RC6mqXtJy-- From pgsql-performance-owner@postgresql.org Thu Nov 20 11:08:06 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 420D2D1B510; Thu, 20 Nov 2003 15:08:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 43766-09; Thu, 20 Nov 2003 11:07:35 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id DC390D1B4F4; Thu, 20 Nov 2003 11:07:30 -0400 (AST) Received: from daithan.intranet.pspl.co.in shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Thu, 20 Nov 2003 08:07:33 -0700 From: Shridhar Daithankar To: Shridhar Daithankar , "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? Date: Thu, 20 Nov 2003 20:37:13 +0530 User-Agent: KMail/1.5.4 Cc: Josh Berkus , pgsql-performance@postgresql.org, PostgreSQL-development References: <200311181558.45847.josh@agliodbs.com> <3FBCD013.3030402@zeut.net> <200311202029.47351.shridhar_daithankar@myrealbox.com> In-Reply-To: <200311202029.47351.shridhar_daithankar@myrealbox.com> MIME-Version: 1.0 Content-Type: Multipart/Mixed; boundary="Boundary-00=_hiNv//Rg/H4BNFT" Message-Id: <200311202037.13091.shridhar_daithankar@myrealbox.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/174 X-Sequence-Number: 4734 --Boundary-00=_hiNv//Rg/H4BNFT Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline On Thursday 20 November 2003 20:29, Shridhar Daithankar wrote: > On Thursday 20 November 2003 20:00, Matthew T. O'Connor wrote: > > Shridhar Daithankar wrote: > > > I will submit a patch that would account deletes in analyze threshold. > > > Since you want to delay the analyze, I would calculate analyze count as > > > > deletes are already accounted for in the analyze threshold. > > Yes. My bad. Deletes are not accounted in initializing analyze count but > later they are used. > > > > I am still wary of inverting vacuum analyze frequency. You think it is > > > better to set inverted default rather than documenting it? > > > > I think inverting the vacuum and analyze frequency is wrong. > > Me. Too. ATM all I can think of this patch attached. Josh, is it sufficient > for you?..:-) use this one. A warning added for too aggressive vacuumming. If it is OK by everybody, we can send it to patches list. Shridhar --Boundary-00=_hiNv//Rg/H4BNFT Content-Type: text/x-diff; charset="iso-8859-1"; name="difffile" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="difffile" *** README.pg_autovacuum.orig Thu Nov 20 19:58:29 2003 --- README.pg_autovacuum Thu Nov 20 20:35:34 2003 *************** *** 141,163 **** depending on the mixture of table activity (insert, update, or delete): ! - If the number of (inserts + updates + deletes) > AnalyzeThreshold, then only an analyze is performed. ! - If the number of (deletes + updates) > VacuumThreshold, then a vacuum analyze is performed. VacuumThreshold is equal to: ! vacuum_base_value + (vacuum_scaling_factor * "number of tuples in the table") ! AnalyzeThreshold is equal to: ! analyze_base_value + (analyze_scaling_factor * "number of tuples in the table") ! The AnalyzeThreshold defaults to half of the VacuumThreshold since it represents a much less expensive operation (approx 5%-10% of vacuum), and running ANALYZE more often should not substantially degrade system performance. Sleeping: --------- --- 141,191 ---- depending on the mixture of table activity (insert, update, or delete): ! - If the number of (inserts + updates + deletes) >= AnalyzeThreshold, then only an analyze is performed. ! - If the number of (deletes + updates) >= VacuumThreshold, then a vacuum analyze is performed. VacuumThreshold is equal to: ! vacuum_base_value + (vacuum_scaling_factor * "number of tuples in the ! table") AnalyzeThreshold is equal to: ! analyze_base_value + (analyze_scaling_factor * "number of tuples in the ! table") The AnalyzeThreshold defaults to half of the VacuumThreshold since it represents a much less expensive operation (approx 5%-10% of vacuum), and running ANALYZE more often should not substantially degrade system performance. + Examples: + + Following table shows typical usage of pg_autovacuum settings. + These are put here so that a DBA can have some starting point while + tuning pg_autovacuum. + + Vacuum is triggered by updates and deletes. So in case of vacuum, + last column indicates total of updates and deletes required + to trigger vacuum. In case of analyze, the operations would count total + number of inserts, updates and deletes. + + Base Scaling factor Records No. of Operations + 1,000 1 10,000 11,000 + 1,000 2 10,000 21,000 + 1,000 0.5 10,000 6,000 + 1,000 0.1 10,000 2,000 + + Although analyze is cheaper operation compared to vacuum, + it might be needed less often. The default is to analyze twice as much as + vacuum but that might be too aggressive for some installations. It is advised + thatsuch installation tune their analyze threshold separately, rather than + relying upon the default behaviour. + + Furthermore, for aggressive vacuum/analyze behaviour, it is recommended that + scaling factor is set to less than 1. However too aggresive operation can affect + performance of normal database operations adversely. Do not apply such setting + to production databases without prior testing. + Sleeping: --------- --Boundary-00=_hiNv//Rg/H4BNFT-- From pgsql-hackers-owner@postgresql.org Thu Nov 20 12:33:46 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DC2D3D1CA5C; Thu, 20 Nov 2003 16:33:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 64936-02; Thu, 20 Nov 2003 12:32:57 -0400 (AST) Received: from mta9.srv.hcvlny.cv.net (mta9.srv.hcvlny.cv.net [167.206.5.42]) by svr1.postgresql.org (Postfix) with ESMTP id 84379D1C991; Thu, 20 Nov 2003 12:32:38 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta9.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HON00D7QSLWVA@mta9.srv.hcvlny.cv.net>; Thu, 20 Nov 2003 11:32:22 -0500 (EST) Date: Thu, 20 Nov 2003 11:32:24 -0500 From: "Matthew T. O'Connor" Subject: Re: [PERFORM] More detail on settings for pgavd? In-reply-to: <200311202029.47351.shridhar_daithankar@myrealbox.com> To: Shridhar Daithankar Cc: Josh Berkus , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBCEC98.4080309@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=us-ascii; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <200311181558.45847.josh@agliodbs.com> <3FBC6BED.9090809@myrealbox.com> <3FBCD013.3030402@zeut.net> <200311202029.47351.shridhar_daithankar@myrealbox.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/1130 X-Sequence-Number: 47418 Shridhar Daithankar wrote: >On Thursday 20 November 2003 20:00, Matthew T. O'Connor wrote: > > >>Shridhar Daithankar wrote: >> >> >>>I am still wary of inverting vacuum analyze frequency. You think it is >>>better to set inverted default rather than documenting it? >>> >>> >>I think inverting the vacuum and analyze frequency is wrong. >> >> >Me. Too. ATM all I can think of this patch attached. Josh, is it sufficient >for you?..:-) > > The patch just adds an example to the README, this looks ok to me. >Matthew, I am confyused about one thing. Why would autovacuum count updates >while checking for analyze threshold? Analyze does not change statistics >right? ( w.r.t line 1072, pg_autovacuum.c). For updating statistics, only >inserts+deletes should suffice, isn't it? > > An update is the equivelant of an insert and a delete, so it counts towards the analyze count as much as an insert. >Other than that, I think autovacuum does everything it can. > > It could be more customizable. From pgsql-performance-owner@postgresql.org Thu Nov 20 13:19:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 20DEAD1B578; Thu, 20 Nov 2003 17:19:44 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 66727-10; Thu, 20 Nov 2003 13:19:12 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 4B4B1D1C9DE; Thu, 20 Nov 2003 13:19:11 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3950080; Thu, 20 Nov 2003 09:19:41 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: "Matthew T. O'Connor" , Shridhar Daithankar Subject: Re: [HACKERS] More detail on settings for pgavd? Date: Thu, 20 Nov 2003 09:18:30 -0800 User-Agent: KMail/1.4.3 Cc: pgsql-performance@postgresql.org, PostgreSQL-development References: <200311181558.45847.josh@agliodbs.com> <3FBC6BED.9090809@myrealbox.com> <3FBCD013.3030402@zeut.net> In-Reply-To: <3FBCD013.3030402@zeut.net> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311200918.30838.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/175 X-Sequence-Number: 4735 Matthew, > For small tables, you don't need to vacuum too often. In the testing I > did a small table ~100 rows, didn't really show significant performance > degredation until it had close to 1000 updates. This is accounted for by using the "threshold" value. That way small tables get vacuumed less often. However, the way large tables work is very different and I think your strategy shows a lack of testing on large active tables. > For large tables, > vacuum is so expensive, that you don't want to do it very often, and > scanning the whole table when there is only 5% wasted space is not very > helpful. 5% is probably too low, you're right ... in my experience, performance degredation starts to set in a 10-15% updates to, for example, a 1.1 million row table, particularly since users tend to request the most recently updated rows. As long as we have the I/O issues that Background Writer and ARC are intended to solve, though, I can see being less agressive on the defaults; perhaps 20% or 25%. If you wait until 110% of a 1.1 million row table is updated, though, that vaccuum will take an hour or more. Additionally, you are not thinking of this in terms of an overall database maintanence strategy. Lazy Vacuum needs to stay below the threshold of the Free Space Map (max_fsm_pages) to prevent creeping bloat from setting in to your databases. With proper configuration of pg_avd, vacuum_mem and FSM values, it should be possible to never run a VACUUM FULL again, and as of 7.4 never run an REINDEX again either. But this means running vacuum frequently enough that your max_fsm_pages threshold is never reached. Which for a large database is going to have to be more frequently than 110% updates, because setting 20,000,000 max_fsm_pages will eat your RAM. > Yes, the I set the defaults a little high perhaps so as to err on the > side of caution. I didn't want people to say pg_autovacuum kills the > performance of my server. A small table will get vacuumed, just not > until it has reached the threshold. So a table with 100 rows, will get > vacuumed after 1200 updates / deletes. Ok, I can see that for small tables. > In my testing it showed that > there was no major performance problems until you reached several > thousand updates / deletes. Sure. But several thousand updates can be only 2% of a very large table. > HUH? analyze is very very cheap compared to vacuum. Why not do it more > often? Because nothing is cheap if it's not needed. Analyze is needed only as often as the *aggregate distribution* of data in the tables changes. Depending on the application, this could be frequently, but far more often (in my experience running multiple databases for several clients) the data distribution of very large tables changes very slowly over time. One client's database, for example, that I have running VACUUM on chron scripts runs on this schedule for the main tables: VACUUM only: twice per hour VACUUM ANALYZE: twice per day On the other hand, I've another client's database where most activity involves updates to entire classes of records. They run ANALYZE at the end of every transaction. So if you're going to have a seperate ANALYZE schedule at all, it should be slightly less frequent than VACUUM for large tables. Either that, or drop the idea, and simplify pg_avd by running VACUUM ANALYZE all the time instead of having 2 seperate schedules. BUT .... now I see how you arrived at the logic you did. If you're testing only on small tables, and not vacuuming them until they reach 110% updates, then you *would* need to analyze more frequently. This is because of your threshold value ... you'd want to analyze the small table as soon as even 30% of its rows changed. So the answer is to dramatically lower the threshold for the small tables. > What I think I am hearing is that people would like very much to be able > to tweak the settings of pg_autovacuum for individual tables / databases > etc. Not from me you're not. Though that would be nice, too. So, my suggested defaults based on our conversation above: Vacuum threshold: 1000 records Vacuum scale factor: 0.2 Analyze threshold: 50 records Analyze scale factor: 0.3 -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Thu Nov 20 14:01:32 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 79B0ED1B545; Thu, 20 Nov 2003 18:01:31 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78548-04; Thu, 20 Nov 2003 14:01:01 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 1BCAFD1C9E5; Thu, 20 Nov 2003 14:01:00 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3950317; Thu, 20 Nov 2003 10:01:45 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: Shridhar Daithankar Subject: Re: More detail on settings for pgavd? Date: Thu, 20 Nov 2003 10:00:34 -0800 User-Agent: KMail/1.4.3 Cc: pgsql-performance@postgresql.org, PostgreSQL-development References: <200311181558.45847.josh@agliodbs.com> <200311190906.15828.josh@agliodbs.com> <3FBC6BED.9090809@myrealbox.com> In-Reply-To: <3FBC6BED.9090809@myrealbox.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311201000.34694.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/176 X-Sequence-Number: 4736 Shridhar, > I would say -V 0.2-0.4 could be great as well. Fact to emphasize is that > thresholds less than 1 should be used. Yes, but not thresholds, scale factors of less than 1.0. Thresholds should still be in the range of 100 to 1000. > I will submit a patch that would account deletes in analyze threshold. > Since you want to delay the analyze, I would calculate analyze count as > > n=updates + inserts *-* deletes I'm not clear on how this is a benefit. Deletes affect the statistics, too. > What does statistics gather BTW? Just number of rows or something else as > well? I think I would put that on Hackers separately. Number of tuples, degree of uniqueness, some sample values, and high/low values. Just query your pg_statistics view for an example. > I am still wary of inverting vacuum analyze frequency. You think it is > better to set inverted default rather than documenting it? See my post to Matthew. -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Thu Nov 20 14:49:12 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C7CD9D1B584; Thu, 20 Nov 2003 18:49:09 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 88664-02; Thu, 20 Nov 2003 14:48:39 -0400 (AST) Received: from arbor.net (division.aa.arbor.net [204.181.64.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5DA0BD1B4F4; Thu, 20 Nov 2003 14:48:37 -0400 (AST) Received: by arbor.net (Postfix, from userid 1065) id A03442A891; Thu, 20 Nov 2003 13:48:21 -0500 (EST) Received: from localhost (localhost [127.0.0.1]) by arbor.net (Postfix) with ESMTP id 93DF12A88C; Thu, 20 Nov 2003 13:48:21 -0500 (EST) Date: Thu, 20 Nov 2003 13:48:21 -0500 (EST) From: Chester Kustarz To: Josh Berkus Cc: pgsql-performance@postgresql.org, PostgreSQL-development Subject: Re: [HACKERS] More detail on settings for pgavd? In-Reply-To: <200311200918.30838.josh@agliodbs.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/177 X-Sequence-Number: 4737 On Thu, 20 Nov 2003, Josh Berkus wrote: > Additionally, you are not thinking of this in terms of an overall database > maintanence strategy. Lazy Vacuum needs to stay below the threshold of the > Free Space Map (max_fsm_pages) to prevent creeping bloat from setting in to > your databases. With proper configuration of pg_avd, vacuum_mem and FSM > values, it should be possible to never run a VACUUM FULL again, and as of 7.4 > never run an REINDEX again either. is there any command you can run to see how much of the FSM is filled? is there any way to tell which tables are filling it? > Analyze is needed only as often as the *aggregate distribution* of data in the > tables changes. Depending on the application, this could be frequently, but > far more often (in my experience running multiple databases for several > clients) the data distribution of very large tables changes very slowly over > time. analyze does 2 things for me: 1. gets reasonable aggregate statistics 2. generates STATISTICS # of bins for the most frequent hitters (2) is very important for me. my values typically seem to have power-law like distributions. i need enough bins to reach a "cross-over" point where the last bin is frequent enough to make an index scan useful. also, i want enough bins so that the planner can choose index a or b for: select * from foo where a=n and b=m; the selectivity of either index depends not only on the average selectivity of index a or index b, but on n and m as well. for example, 1M row table: value % of rows v1 23 v2 12 v3 4.5 v4 4 v5 3.5 ... you can see that picking an index for =v1 would be poor. picking the 20th most common value would be 0.5% selective. much better. of course this breaks down for more complex operators, but = is fairly common. > So if you're going to have a seperate ANALYZE schedule at all, it should be > slightly less frequent than VACUUM for large tables. Either that, or drop > the idea, and simplify pg_avd by running VACUUM ANALYZE all the time instead > of having 2 seperate schedules. i have some tables which are insert only. i do not want to vacuum them because there are never any dead tuples in them and the vacuum grows the indexes. plus it is very expensive (they tables grow rather large.) after they expire i drop the whole table to make room for a newer one (making sort of a rolling log with many large tables.) i need to analyze them every so often so that the planner knows that there is 1 row, 100 rows, 100k rows, 1M. the funny thing is that because i never vacuum the tables, the relpages on the index never grows. don't know if this affects anything (this is on 7.2.3). vacuum is to reclaim dead tuples. this means it depends on update and delete. analyze depends on data values/distribution. this means it depends on insert, update, and delete. thus the dependencies are slightly different between the 2 operations, an so you can come up with use-cases that justify running either more frequently. i am not sure how failed transactions fit into this though, not that i think anybody ever has very many. maybe big rollbacks during testing? From pgsql-performance-owner@postgresql.org Thu Nov 20 15:21:50 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A7C53D1CCE0; Thu, 20 Nov 2003 19:21:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 95691-05; Thu, 20 Nov 2003 15:21:16 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 08848D1D163; Thu, 20 Nov 2003 15:20:59 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAKJKO19006001; Thu, 20 Nov 2003 14:20:24 -0500 (EST) To: Chester Kustarz Cc: Josh Berkus , pgsql-performance@postgresql.org, PostgreSQL-development Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: References: Comments: In-reply-to Chester Kustarz message dated "Thu, 20 Nov 2003 13:48:21 -0500" Date: Thu, 20 Nov 2003 14:20:24 -0500 Message-ID: <6000.1069356024@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/178 X-Sequence-Number: 4738 Chester Kustarz writes: > i have some tables which are insert only. i do not want to vacuum them > because there are never any dead tuples in them and the vacuum grows the > indexes. Those claims cannot both be true. In any case, plain vacuum cannot grow the indexes --- only a VACUUM FULL that moves a significant number of rows could cause index growth. > vacuum is to reclaim dead tuples. this means it depends on update and > delete. analyze depends on data values/distribution. this means it depends on > insert, update, and delete. thus the dependencies are slightly different > between the 2 operations, an so you can come up with use-cases that > justify running either more frequently. Agreed. regards, tom lane From pgsql-performance-owner@postgresql.org Thu Nov 20 16:55:10 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A26CAD1C969; Thu, 20 Nov 2003 20:55:09 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 08196-09; Thu, 20 Nov 2003 16:54:40 -0400 (AST) Received: from arbor.net (division.aa.arbor.net [204.181.64.2]) by svr1.postgresql.org (Postfix) with ESMTP id C9A6DD1B55F; Thu, 20 Nov 2003 16:54:38 -0400 (AST) Received: by arbor.net (Postfix, from userid 1065) id 9E3EE2A891; Thu, 20 Nov 2003 15:54:24 -0500 (EST) Received: from localhost (localhost [127.0.0.1]) by arbor.net (Postfix) with ESMTP id 913352A88C; Thu, 20 Nov 2003 15:54:24 -0500 (EST) Date: Thu, 20 Nov 2003 15:54:24 -0500 (EST) From: Chester Kustarz To: Tom Lane Cc: Josh Berkus , , PostgreSQL-development Subject: Re: [HACKERS] More detail on settings for pgavd? In-Reply-To: <6000.1069356024@sss.pgh.pa.us> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/179 X-Sequence-Number: 4739 On Thu, 20 Nov 2003, Tom Lane wrote: > Those claims cannot both be true. In any case, plain vacuum cannot grow > the indexes --- only a VACUUM FULL that moves a significant number of > rows could cause index growth. er, yeah. you're right of course. having flashbacks of vacuum full. From pgsql-performance-owner@postgresql.org Mon Nov 24 14:56:29 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E8001D1D4BE for ; Thu, 20 Nov 2003 21:05:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 09813-06 for ; Thu, 20 Nov 2003 17:04:30 -0400 (AST) Received: from almaden.ibm.com (p1.almaden.ibm.com [198.4.83.52]) by svr1.postgresql.org (Postfix) with ESMTP id 3EADFD1C950 for ; Thu, 20 Nov 2003 17:04:28 -0400 (AST) Received: from user.almaden.ibm.com (user.almaden.ibm.com [9.1.25.220]) by almaden.ibm.com (AIX4.3/8.9.3p2/8.9.3) with ESMTP id NAA32396 for ; Thu, 20 Nov 2003 13:04:29 -0800 Received: from almaden.ibm.com (sfarrell@yoyo.almaden.ibm.com [9.1.24.63]) by user.almaden.ibm.com (8.12.3/8.12.3/Debian -4) with ESMTP id hAKL4TlZ023952 for ; Thu, 20 Nov 2003 13:04:29 -0800 Message-ID: <3FBD2C5D.8000706@almaden.ibm.com> Date: Thu, 20 Nov 2003 13:04:29 -0800 From: stephen farrell User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031030 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Problem with insert into select... Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/221 X-Sequence-Number: 4781 I'm having a problem with a queyr like: INSERT INTO FACT (x,x,x,x,x,x) SELECT a.key,b.key,c.key,d.key,e.key,f.key from x,a,b,c,d,e,f where x=a and x=b .... -- postgres7.4 is running out of memory. I'm not sure why this would happen -- does it buffer the subselect before doing the insert? Things are pretty big scale: 3gb ram, 32768 shared buffers, 700gb disk, millions of rows in the tables. From pgsql-performance-owner@postgresql.org Thu Nov 20 17:10:06 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1C393D1D570 for ; Thu, 20 Nov 2003 21:10:05 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 09813-08 for ; Thu, 20 Nov 2003 17:09:35 -0400 (AST) Received: from almaden.ibm.com (p1.almaden.ibm.com [198.4.83.52]) by svr1.postgresql.org (Postfix) with ESMTP id C61DDD1D4CE for ; Thu, 20 Nov 2003 17:09:30 -0400 (AST) Received: from user.almaden.ibm.com (user.almaden.ibm.com [9.1.25.220]) by almaden.ibm.com (AIX4.3/8.9.3p2/8.9.3) with ESMTP id NAA50960 for ; Thu, 20 Nov 2003 13:09:31 -0800 Received: from almaden.ibm.com (sfarrell@yoyo.almaden.ibm.com [9.1.24.63]) by user.almaden.ibm.com (8.12.3/8.12.3/Debian -4) with ESMTP id hAKL9VlZ024046 for ; Thu, 20 Nov 2003 13:09:31 -0800 Message-ID: <3FBD2D8B.8020509@almaden.ibm.com> Date: Thu, 20 Nov 2003 13:09:31 -0800 From: stephen farrell User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031030 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Problem with insert into select... Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/180 X-Sequence-Number: 4740 I'm having a problem with a queyr like: INSERT INTO FACT (x,x,x,x,x,x) SELECT a.key,b.key,c.key,d.key,e.key,f.key from x,a,b,c,d,e,f where x=a and x=b .... -- postgres7.4 is running out of memory. I'm not sure why this would happen -- does it buffer the subselect before doing the insert? Things are pretty big scale: 3gb ram, 32768 shared buffers, 700gb disk, millions of rows in the tables. From pgsql-performance-owner@postgresql.org Thu Nov 20 18:59:27 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EA1C2D1B4E6 for ; Thu, 20 Nov 2003 22:59:23 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 30763-01 for ; Thu, 20 Nov 2003 18:58:48 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id F20CED1B4F4 for ; Thu, 20 Nov 2003 18:58:45 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAKMwe19007440; Thu, 20 Nov 2003 17:58:40 -0500 (EST) To: stephen farrell Cc: pgsql-performance@postgresql.org Subject: Re: Problem with insert into select... In-reply-to: <3FBD2D8B.8020509@almaden.ibm.com> References: <3FBD2D8B.8020509@almaden.ibm.com> Comments: In-reply-to stephen farrell message dated "Thu, 20 Nov 2003 13:09:31 -0800" Date: Thu, 20 Nov 2003 17:58:40 -0500 Message-ID: <7439.1069369120@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/181 X-Sequence-Number: 4741 stephen farrell writes: > I'm having a problem with a queyr like: INSERT INTO FACT (x,x,x,x,x,x) > SELECT a.key,b.key,c.key,d.key,e.key,f.key from x,a,b,c,d,e,f where x=a > and x=b .... -- postgres7.4 is running out of memory. I'm not sure > why this would happen -- does it buffer the subselect before doing the > insert? What does EXPLAIN show for the query? And we need to see the exact query and table definitions, not abstractions. regards, tom lane From pgsql-performance-owner@postgresql.org Mon Nov 24 14:56:25 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E77C5D1B54C; Thu, 20 Nov 2003 23:19:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 32640-08; Thu, 20 Nov 2003 19:18:43 -0400 (AST) Received: from mta1.srv.hcvlny.cv.net (mta1.srv.hcvlny.cv.net [167.206.5.67]) by svr1.postgresql.org (Postfix) with ESMTP id 403E4D1B520; Thu, 20 Nov 2003 19:18:40 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta1.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HOO004BHBF8UT@mta1.srv.hcvlny.cv.net>; Thu, 20 Nov 2003 18:18:49 -0500 (EST) Date: Thu, 20 Nov 2003 18:18:27 -0500 From: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: <6000.1069356024@sss.pgh.pa.us> To: Tom Lane Cc: Chester Kustarz , Josh Berkus , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBD4BC3.8030003@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <6000.1069356024@sss.pgh.pa.us> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/220 X-Sequence-Number: 4780 Tom Lane wrote: >Chester Kustarz writes: > > >>vacuum is to reclaim dead tuples. this means it depends on update and >>delete. analyze depends on data values/distribution. this means it depends on >>insert, update, and delete. thus the dependencies are slightly different >>between the 2 operations, an so you can come up with use-cases that >>justify running either more frequently. >> >> >Agreed. > > And that is why pg_autovacuum looks at insert, update and delete when deciding to do an analyze, but only looks at update and delete when deciding to do a vacuum. In addition, this is why pg_autovacuum was given knobs so that the vacuum and analyze thresholds can be set independently. Matthew From pgsql-performance-owner@postgresql.org Thu Nov 20 20:17:48 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 15A02D1D4CE for ; Fri, 21 Nov 2003 00:17:47 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 44878-01 for ; Thu, 20 Nov 2003 20:17:18 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 061F0D1B54E for ; Thu, 20 Nov 2003 20:17:15 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAL0H119007949; Thu, 20 Nov 2003 19:17:02 -0500 (EST) To: Ryszard Lach Cc: pgsql-performance@postgresql.org, Bruce Momjian Subject: Re: duration logging setting in 7.4 In-reply-to: <20031118091646.GA8223@siaco.id.pl> References: <20031117222058.GD3248@siaco.id.pl> <200311180237.hAI2b7s29417@candle.pha.pa.us> <20031118091646.GA8223@siaco.id.pl> Comments: In-reply-to Ryszard Lach message dated "Tue, 18 Nov 2003 10:16:46 +0100" Date: Thu, 20 Nov 2003 19:17:01 -0500 Message-ID: <7948.1069373821@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/182 X-Sequence-Number: 4742 Ryszard Lach writes: > Nov 18 10:05:20 postgres[1348]: [318-1] LOG: duration: 0.297 ms statement: > Nov 18 10:05:20 postgres[1311]: [5477-1] LOG: duration: 0.617 ms statement: > Nov 18 10:05:20 postgres[1312]: [5134-1] LOG: duration: 0.477 ms statement: > Nov 18 10:05:20 postgres[1349]: [318-1] LOG: duration: 0.215 ms statement: > Nov 18 10:05:20 postgres[1313]: [5449-1] LOG: duration: 0.512 ms statement: > Nov 18 10:05:20 postgres[1314]: [5534-1] LOG: duration: 0.420 ms statement: > Nov 18 10:05:20 postgres[1330]: [772-1] LOG: duration: 1.386 ms statement: SELECT * FROM mytablemius WHERE id = 0; > Nov 18 10:05:20 postgres[1315]: [5757-1] LOG: duration: 0.417 ms statement: > Nov 18 10:05:20 postgres[1316]: [5885-1] LOG: duration: 0.315 ms statement: > Nov 18 10:05:20 postgres[1317]: [5914-1] LOG: duration: 0.301 ms statement: > Nov 18 10:05:20 postgres[1318]: [5990-1] LOG: duration: 0.293 ms statement: > Nov 18 10:05:20 postgres[1319]: [6009-1] LOG: duration: 0.211 ms statement: > Nov 18 10:05:20 postgres[1320]: [6039-1] LOG: duration: 0.188 ms statement: Is it possible that you're sending a lot of queries that have an initial newline in the text? I'd expect the first line of log output for such a query to look as above. regards, tom lane From pgsql-performance-owner@postgresql.org Mon Nov 24 14:56:22 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9DD7FD1B520; Fri, 21 Nov 2003 00:41:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 49229-04; Thu, 20 Nov 2003 20:41:05 -0400 (AST) Received: from mta2.srv.hcvlny.cv.net (mta2.srv.hcvlny.cv.net [167.206.5.68]) by svr1.postgresql.org (Postfix) with ESMTP id 81586D1B50A; Thu, 20 Nov 2003 20:41:02 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta2.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HOO001ATF7S9Y@mta2.srv.hcvlny.cv.net>; Thu, 20 Nov 2003 19:40:45 -0500 (EST) Date: Thu, 20 Nov 2003 19:40:15 -0500 From: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: <200311200918.30838.josh@agliodbs.com> To: Josh Berkus Cc: Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBD5EEF.7060205@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <200311181558.45847.josh@agliodbs.com> <3FBC6BED.9090809@myrealbox.com> <3FBCD013.3030402@zeut.net> <200311200918.30838.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/219 X-Sequence-Number: 4779 Josh Berkus wrote: >Matthew, > > >>For small tables, you don't need to vacuum too often. In the testing I >>did a small table ~100 rows, didn't really show significant performance >>degredation until it had close to 1000 updates. >> >> >This is accounted for by using the "threshold" value. That way small tables >get vacuumed less often. However, the way large tables work is very different >and I think your strategy shows a lack of testing on large active tables. > > Probably more true than I would like to think... >>For large tables, >>vacuum is so expensive, that you don't want to do it very often, and >>scanning the whole table when there is only 5% wasted space is not very >>helpful. >> >> >5% is probably too low, you're right ... in my experience, performance >degredation starts to set in a 10-15% updates to, for example, a 1.1 million >row table, particularly since users tend to request the most recently updated >rows. As long as we have the I/O issues that Background Writer and ARC are >intended to solve, though, I can see being less agressive on the defaults; >perhaps 20% or 25%. If you wait until 110% of a 1.1 million row table is >updated, though, that vaccuum will take an hour or more. > > True, but I think it would be one hour once, rather than 30 minutes 4 times. >Additionally, you are not thinking of this in terms of an overall database >maintanence strategy. Lazy Vacuum needs to stay below the threshold of the >Free Space Map (max_fsm_pages) to prevent creeping bloat from setting in to >your databases. With proper configuration of pg_avd, vacuum_mem and FSM >values, it should be possible to never run a VACUUM FULL again, and as of 7.4 >never run an REINDEX again either. > > This is one of the things I had hoped to add to pg_autovacuum, but never got to. In addition to just the information from the stats collector on inserts updates and deletes, pg_autovacuum should also look at the FSM, and make decisions based on it. Anyone looking for a project? >But this means running vacuum frequently enough that your max_fsm_pages >threshold is never reached. Which for a large database is going to have to >be more frequently than 110% updates, because setting 20,000,000 >max_fsm_pages will eat your RAM. > > Again, the think the only way to do this efficiently is to look at the FSM. Otherwise the only way to make sure you keep the FSM populated is to run vacuum more than needed. >>Yes, the I set the defaults a little high perhaps so as to err on the >>side of caution. I didn't want people to say pg_autovacuum kills the >>performance of my server. A small table will get vacuumed, just not >>until it has reached the threshold. So a table with 100 rows, will get >>vacuumed after 1200 updates / deletes. >> >> >Ok, I can see that for small tables. > > >>In my testing it showed that >>there was no major performance problems until you reached several >>thousand updates / deletes. >> >> >Sure. But several thousand updates can be only 2% of a very large table. > > But I can't imagine that 2% makes any difference on a large table. In fact I would think that 10-15% would hardly be noticable, beyond that I'm not sure. >>HUH? analyze is very very cheap compared to vacuum. Why not do it more >>often? >> >> >Because nothing is cheap if it's not needed. > >Analyze is needed only as often as the *aggregate distribution* of data in the >tables changes. Depending on the application, this could be frequently, but >far more often (in my experience running multiple databases for several >clients) the data distribution of very large tables changes very slowly over >time. > > Valid points, and again I think this points to the fact that pg_autovacuum needs to be more configurable. Being able to set different thresholds for different tables will help considerably. In fact, you may find that some tables should have a vac threshold much larger than the analyze thresold, while other tables might want the opposite. >One client's database, for example, that I have running VACUUM on chron >scripts runs on this schedule for the main tables: >VACUUM only: twice per hour >VACUUM ANALYZE: twice per day > > I would be surprized if you can notice the difference between a vacuum analyze and a vacuum, especially on large tables. >On the other hand, I've another client's database where most activity involves >updates to entire classes of records. They run ANALYZE at the end of every >transaction. > >So if you're going to have a seperate ANALYZE schedule at all, it should be >slightly less frequent than VACUUM for large tables. Either that, or drop >the idea, and simplify pg_avd by running VACUUM ANALYZE all the time instead >of having 2 seperate schedules. > > I think you need two separate schedules. There are lots of times where a vacuum doesn't help, and an analyze is all that is needed, and an analyze is MUCH cheaper than a vacuum. >BUT .... now I see how you arrived at the logic you did. If you're testing >only on small tables, and not vacuuming them until they reach 110% updates, >then you *would* need to analyze more frequently. This is because of your >threshold value ... you'd want to analyze the small table as soon as even 30% >of its rows changed. > >So the answer is to dramatically lower the threshold for the small tables. > > Perhaps. >>What I think I am hearing is that people would like very much to be able >>to tweak the settings of pg_autovacuum for individual tables / databases >>etc. >> >> >Not from me you're not. Though that would be nice, too. > >So, my suggested defaults based on our conversation above: > >Vacuum threshold: 1000 records >Vacuum scale factor: 0.2 >Analyze threshold: 50 records >Analyze scale factor: 0.3 > > I'm open to discussion on changing the defaults. Perhaps what it would be better to use some non-linear (perhaps logorithmic) scaling factor. So that you wound up with something roughly like this: #tuples activity% for vacuum 1k 100% 10k 70% 100k 45% 1M 20% 10M 10% 100M 8% Thanks for the lucid feedback / discussion. autovacuum is a feature that, despite it's simple implementation, has generated a lot of feedback from users, and I would really like to see it become something closer to what it should be. From pgsql-performance-owner@postgresql.org Thu Nov 20 22:05:45 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (unknown [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DECCED1B4E6 for ; Fri, 21 Nov 2003 02:05:41 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 66594-02 for ; Thu, 20 Nov 2003 22:05:14 -0400 (AST) Received: from almaden.ibm.com (p1.almaden.ibm.com [198.4.83.52]) by svr1.postgresql.org (Postfix) with ESMTP id 76938D1B54E for ; Thu, 20 Nov 2003 22:05:10 -0400 (AST) Received: from user.almaden.ibm.com (user.almaden.ibm.com [9.1.25.220]) by almaden.ibm.com (AIX4.3/8.9.3p2/8.9.3) with ESMTP id SAA58840 for ; Thu, 20 Nov 2003 18:05:13 -0800 Received: from almaden.ibm.com (sfarrell@yoyo.almaden.ibm.com [9.1.24.63]) by user.almaden.ibm.com (8.12.3/8.12.3/Debian -4) with ESMTP id hAL259lZ027038 for ; Thu, 20 Nov 2003 18:05:12 -0800 Message-ID: <3FBD72D5.1040202@almaden.ibm.com> Date: Thu, 20 Nov 2003 18:05:09 -0800 From: stephen farrell User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031030 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Re: Problem with insert into select... References: <3FBD2D8B.8020509@almaden.ibm.com> <7439.1069369120@sss.pgh.pa.us> In-Reply-To: <7439.1069369120@sss.pgh.pa.us> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/183 X-Sequence-Number: 4743 Ok -- so we created indexes and it was able to complete successfully. But why would creating indexes affect the memory footprint, and should it? Does it buffer the sub-select before doing the insert, or does it do the insert record-by-record? See correspondence below for details: Steve, With the indexes created it worked. It took about 4 hours, but it inserted all of the records. stephen farrell 11/20/2003 05:22 PM To: James Rhodes/Almaden/IBM@IBMUS cc: Subject: Re: [Fwd: Re: [PERFORM] Problem with insert into select...] if you do "explain" before the sql statement (e.g., "explain select * from foo"), it'll tell you the query plan. James Rhodes wrote: > > Steve, > > Here is the detailed structure of the tables and the query that > is failing (the "INSERT INTO FACT" query) and I attached the logfile. > Also what is EXPLAIN??? > > CREATE TABLE RAW ( RAW_KEY serial, PATNO_TEXT VARCHAR (9), > APPDATE_DATETIME VARCHAR (11), ISDATE_DATETIME VARCHAR (11), > WHATEVERSNO_TEXT VARCHAR (5), WHATEVERSNO_NUMBER VARCHAR (6), APPNO_TEXT > VARCHAR (10), TITLE_TEXT TEXT, USCLASS_TEXT VARCHAR (14), > USCLASS_TEXTLIST_TEXT TEXT, AUTHORCODE_TEXT VARCHAR (9), > AUTHORNORM_TEXT VARCHAR (195), AUTHOR_TEXT VARCHAR (212), > AUTHOR_TEXTLIST_TEXT TEXT, AUTHORADDRESS_TEXT VARCHAR (84), > AUTHORADDRESS_TEXTLIST_TEXT TEXT, INVENTOR_TEXT VARCHAR (50), > INVENTOR_TEXTLIST_TEXT TEXT, INVENTORADDRESS_TEXT VARCHAR (90), > INVENTORADDRESS_TEXTLIST_TEXT TEXT, AGENT_TEXT TEXT, AGENT_TEXTLIST_TEXT > TEXT, USSEARCHFIELD_TEXT VARCHAR (26), USSEARCHFIELD_TEXTLIST_TEXT > VARCHAR (150), USREFISDATE_TEXT VARCHAR (13), USREFISDATE_TEXTLIST_TEXT > TEXT, USREFNAME_TEXT VARCHAR (34), USREFNAME_TEXTLIST_TEXT TEXT, > ABSTRACT_TEXT TEXT, ABSTRACT_TEXTLIST_TEXT TEXT, ABSTRACT_RICHTEXT_PAR > TEXT, WHATEVERS_RICHTEXT_PAR TEXT, USREFPATNO_RICHTEXT_PAR TEXT, PRIMARY > KEY(RAW_KEY)); > > > CREATE TABLE ISSUE_TIME ( > TAB_KEY serial, > ISDATE_DATETIME varchar (8), > MONTH INT, > DAY INT, > YEAR INT > , PRIMARY KEY(TAB_KEY)) > > CREATE TABLE SOMETHING_NUMBER ( > TAB_KEY serial, > PATNO_TEXT varchar (7) > , PRIMARY KEY(TAB_KEY)) > > CREATE TABLE APP_TIME ( > TAB_KEY serial, > APPDATE_DATETIME varchar (8), > MONTH INT, > DAY INT, > YEAR INT > , PRIMARY KEY(TAB_KEY)) > > CREATE TABLE AUTHOR ( > TAB_KEY serial, > CODE varchar (6), > AUTHOR text > , PRIMARY KEY(TAB_KEY)) > > CREATE TABLE APPLICATION_NUMBER ( > TAB_KEY serial, > APPNO_TEXT varchar (7) > , PRIMARY KEY(TAB_KEY)) > > CREATE TABLE WHATEVERS ( > TAB_KEY serial, > abstract_richtext_par text, > WHATEVERS_richtext_par text, > raw_key int, > title_text text > , PRIMARY KEY(TAB_KEY)) > > CREATE TABLE FACT (DYN_DIM1 BIGINT, DYN_DIM2 BIGINT,DYN_DIM3 > BIGINT,ISSUE_TIME BIGINT, SOMETHING_NUMBER BIGINT, APP_TIME BIGINT, > AUTHOR BIGINT, APPLICATION_NUMBER BIGINT, WHATEVERS BIGINT) > > INSERT INTO FACT (ISSUE_TIME, SOMETHING_NUMBER, APP_TIME, AUTHOR, > APPLICATION_NUMBER, WHATEVERS) SELECT ISSUE_TIME.TAB_KEY, > SOMETHING_NUMBER.TAB_KEY, APP_TIME.TAB_KEY, AUTHOR.TAB_KEY, > APPLICATION_NUMBER.TAB_KEY, WHATEVERS.TAB_KEY FROM ISSUE_TIME, > SOMETHING_NUMBER, APP_TIME, AUTHOR, APPLICATION_NUMBER, WHATEVERS, raw > WHERE ISSUE_TIME.ISDATE_DATETIME=raw.ISDATE_DATETIME AND > SOMETHING_NUMBER.PATNO_TEXT=raw.PATNO_TEXT AND > APP_TIME.APPDATE_DATETIME=raw.APPDATE_DATETIME AND > AUTHOR.CODE=AUTHORCODE_TEXT AND AUTHOR.AUTHOR=(AUTHOR_TEXT || > ' | ' || AUTHOR_TEXTLIST_TEXT) AND > APPLICATION_NUMBER.APPNO_TEXT=raw.APPNO_TEXT AND > WHATEVERS.raw_key=raw.raw_key Tom Lane wrote: > stephen farrell writes: > >>I'm having a problem with a queyr like: INSERT INTO FACT (x,x,x,x,x,x) >>SELECT a.key,b.key,c.key,d.key,e.key,f.key from x,a,b,c,d,e,f where x=a >>and x=b .... -- postgres7.4 is running out of memory. I'm not sure >>why this would happen -- does it buffer the subselect before doing the >>insert? > > > What does EXPLAIN show for the query? And we need to see the exact > query and table definitions, not abstractions. > > regards, tom lane From pgsql-performance-owner@postgresql.org Fri Nov 21 02:25:51 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D1F48D1D2BB; Fri, 21 Nov 2003 06:25:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 02910-05; Fri, 21 Nov 2003 02:25:15 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 56340D1D247; Fri, 21 Nov 2003 02:25:14 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3953394; Thu, 20 Nov 2003 22:26:00 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? Date: Thu, 20 Nov 2003 22:24:45 -0800 User-Agent: KMail/1.4.3 Cc: Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development References: <200311181558.45847.josh@agliodbs.com> <200311200918.30838.josh@agliodbs.com> <3FBD5EEF.7060205@zeut.net> In-Reply-To: <3FBD5EEF.7060205@zeut.net> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311202224.45065.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/184 X-Sequence-Number: 4744 Matthew, > > 110% of a 1.1 million row table is updated, though, that vaccuum will > > take an hour or more. > > True, but I think it would be one hour once, rather than 30 minutes 4 > times. Well, generally it would be about 6-8 times at 2-4 minutes each. > This is one of the things I had hoped to add to pg_autovacuum, but never > got to. In addition to just the information from the stats collector on > inserts updates and deletes, pg_autovacuum should also look at the FSM, > and make decisions based on it. Anyone looking for a project? Hmmm ... I think that's the wrong approach. Once your database is populated, it's very easy to determine how to set the FSM for a given pg_avd level. If you're vacuuming after 20% updates, for example, just set fsm_pages to 20% of the total database pages plus growth & safety margins. I'd be really reluctant to base pv-avd frequency on the fsm settings instead. What if the user loads 8GB of data but leaves fsm_pages at the default of 10,000? You can't do much with that; you'd have to vacuum if even 1% of the data changed. The other problem is that calculating data pages from a count of updates+deletes would require pg_avd to keep more statistics and do more math for every table. Do we want to do this? > But I can't imagine that 2% makes any difference on a large table. In > fact I would think that 10-15% would hardly be noticable, beyond that > I'm not sure. I've seen performance lag at 10% of records, especially in tables where both update and select activity focus on one subset of the table (calendar tables, for example). > Valid points, and again I think this points to the fact that > pg_autovacuum needs to be more configurable. Being able to set > different thresholds for different tables will help considerably. In > fact, you may find that some tables should have a vac threshold much > larger than the analyze thresold, while other tables might want the > opposite. Sure. Though I think we can make the present configuration work with a little adjustment of the numbers. I'll have a chance to test on production databases soon. > I would be surprized if you can notice the difference between a vacuum > analyze and a vacuum, especially on large tables. It's substantial for tables with high statistics settings. A 1,000,000 row table with 5 columns set to statistics=250 can take 3 minutes to analyze on a medium-grade server. > I think you need two separate schedules. There are lots of times where > a vacuum doesn't help, and an analyze is all that is needed Agreed. And I've just talked to a client who may want to use pg_avd's ANALYZE scheduling but not use vacuum at all. BTW, I think we should have a setting for this; for example, if -V is -1, don't vacuum. > I'm open to discussion on changing the defaults. Perhaps what it would > be better to use some non-linear (perhaps logorithmic) scaling factor. > So that you wound up with something roughly like this: > > #tuples activity% for vacuum > 1k 100% > 10k 70% > 100k 45% > 1M 20% > 10M 10% > 100M 8% That would be cool, too. Though a count of data pages would be a better scale than a count of rows, and equally obtainable from pg_class. > Thanks for the lucid feedback / discussion. autovacuum is a feature > that, despite it's simple implementation, has generated a lot of > feedback from users, and I would really like to see it become something > closer to what it should be. Well, I hope to help now. Until very recently, I've not had a chance to seriously look at pg_avd and test it in production. Now that I do, I'm interested in improving it. -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Fri Nov 21 04:54:01 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9E9BBD1D59A for ; Fri, 21 Nov 2003 08:53:56 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 29079-03 for ; Fri, 21 Nov 2003 04:53:27 -0400 (AST) Received: from nitrogen.id.pl (nitrogen.id.pl [193.178.214.5]) by svr1.postgresql.org (Postfix) with SMTP id A1CE3D1D563 for ; Fri, 21 Nov 2003 04:53:23 -0400 (AST) Received: (qmail 26805 invoked by uid 0); 21 Nov 2003 08:53:21 -0000 Received: from unknown (HELO siaco.id.pl) (213.25.114.8) by smtp.id.pl with SMTP; 21 Nov 2003 08:53:21 -0000 Received: (qmail 749 invoked by uid 1000); 21 Nov 2003 08:53:17 -0000 Date: Fri, 21 Nov 2003 09:53:17 +0100 From: Ryszard Lach To: Tom Lane Cc: pgsql-performance@postgresql.org, Bruce Momjian Subject: Re: duration logging setting in 7.4 Message-ID: <20031121085317.GD331@siaco.id.pl> Reply-To: Ryszard Lach References: <20031117222058.GD3248@siaco.id.pl> <200311180237.hAI2b7s29417@candle.pha.pa.us> <20031118091646.GA8223@siaco.id.pl> <7948.1069373821@sss.pgh.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-2 Content-Disposition: inline In-Reply-To: <7948.1069373821@sss.pgh.pa.us> X-My-GPG-Key: echo | mail -s "send key pub" ryszard@lach.name User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/185 X-Sequence-Number: 4745 On Thu, Nov 20, 2003 at 07:17:01PM -0500, Tom Lane wrote: > > > Is it possible that you're sending a lot of queries that have an initial > newline in the text? I'd expect the first line of log output for such a > query to look as above. I don't think so, but it is possible, that queries have e.g. two semicolons on end - I've just noticed, that separating two queries with two or more semicolons gives one empty log entry for each redundand semicolon. We'll debug our application keeping this in mind. Richard. -- "First they ignore you. Then they laugh at you. Then they fight you. Then you win." - Mohandas Gandhi. From pgsql-performance-owner@postgresql.org Mon Nov 24 14:55:32 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C72B5D1D570 for ; Fri, 21 Nov 2003 10:10:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 44624-04 for ; Fri, 21 Nov 2003 06:10:22 -0400 (AST) Received: from mailhub1.sghms.ac.uk (firewall.sghms.ac.uk [194.82.50.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3B97BD1D695 for ; Fri, 21 Nov 2003 06:10:19 -0400 (AST) Received: from [194.82.51.24] (helo=imail) by mailhub1.sghms.ac.uk with esmtp (Exim 4.24) id 1AN8Eh-0008JU-DF; Fri, 21 Nov 2003 10:09:59 +0000 Received: from [172.16.20.3] (mrc1-003.sghms.ac.uk [172.16.20.3]) by imail.sghms.ac.uk (iPlanet Messaging Server 5.2 Patch 1 (built Aug 19 2002)) with ESMTPA id <0HOP00AD15KNGA@imail.sghms.ac.uk>; Fri, 21 Nov 2003 10:09:59 +0000 (GMT) Date: Fri, 21 Nov 2003 10:08:42 +0000 From: Adam Witney Subject: Re: Hardware advice In-reply-to: To: "scott.marlowe" Cc: pgsql-performance Message-id: MIME-version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT User-Agent: Microsoft-Entourage/10.1.1.2418 X-MailScanner-Information: Please contact sysadmin at sghms.ac.uk for more information X-MailScanner-MH1: Found to be clean X-MailScanner-SpamCheck: not spam, SpamAssassin (score=-4.9, required 5, APARENTLY_FROM_MYSELF 1.00, BAYES_00 -4.90, FROM_SGHMS -1.00) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/218 X-Sequence-Number: 4778 On 30/5/03 6:17 pm, "scott.marlowe" wrote: > On Fri, 30 May 2003, Adam Witney wrote: > >> Hi scott, >> >> Thanks for the info >> >>> You might wanna do something like go to all 146 gig drives, put a mirror >>> set on the first 20 or so gigs for the OS, and then use the remainder >>> (5x120gig or so ) to make your RAID5. The more drives in a RAID5 the >>> better, generally, up to about 8 or 12 as the optimal for most setups. >> >> I am not quite sure I understand what you mean here... Do you mean take 20Gb >> from each of the 5 drives to setup a 20Gb RAID 1 device? Or just from the >> first 2 drives? > > You could do it either way, since the linux kernel supports more than 2 > drives in a mirror. But, this costs on writes, so don't do it for things > like /var or the pg_xlog directory. > > There are a few ways you could arrange 5 146 gig drives. > > One might be to make the first 20 gig on each drive part of a mirror set > where the first two drives are the live mirror, and the next three are hot > spares. Then you could setup your RAID5 to have 4 live drives and 1 hot > spare. > > Hot spares are nice to have because they provide for the shortest period > of time during which your machine is running with a degraded RAID array. > > note that in linux you can set the kernel parameter > dev.raid.speed_limit_max and dev.raid.speed_limit_min to control the > rebuild bandwidth used so that when a disk dies you can set a compromise > between fast rebuilds, and lowering the demands on the I/O subsystem > during a rebuild. The max limit default is 100k / second, which is quite > slow. On a machine with Ultra320 gear, you could set that to 10 ot 20 > megs a second and still not saturate your SCSI buss. > > Now that I think of it, you could probably set it up so that you have a > mirror set for the OS, one for pg_xlog, and then use the rest of the > drives as RAID5. Then grab space on the fifth drive to make a hot spare > for both the pg_xlog and the OS drive. > > Drive 0 > [OS RAID1 20 Gig D0][big data drive RAID5 106 Gig D0] > Drive 1 > [OS RAID1 20 Gig D1][big data drive RAID5 106 Gig D1] > Drive 2 > [pg_xlog RAID1 20 gig D0][big data drive RAID5 106 Gig D2] > Drive 3 > [pg_xlog RAID1 20 gig D1][big data drive RAID5 106 Gig D3] > Drive 4 > [OS hot spare 20 gig][g_clog hot spare 20 gig][big data drive RAID5 106 > Gig hot spare] > > That would give you ~ 300 gigs storage. > > Of course, there will likely be slightly less performance than you might > get from dedicated RAID arrays for each RAID1/RAID5 set, but my guess is > that by having 4 (or 5 if you don't want a hot spare) drives in the RAID5 > it'll still be faster than a dedicated 3 drive RAID array. > Hi Scott, Just following up a post from a few months back... I have now purchased the hardware, do you have a recommended/preferred Linux distro that is easy to configure for software RAID? Thanks again Adam -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. From pgsql-performance-owner@postgresql.org Mon Nov 24 14:55:01 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EC459D1D643 for ; Fri, 21 Nov 2003 10:19:35 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 47972-02 for ; Fri, 21 Nov 2003 06:19:07 -0400 (AST) Received: from nitrogen.id.pl (nitrogen.id.pl [193.178.214.5]) by svr1.postgresql.org (Postfix) with SMTP id 7F8EED1B553 for ; Fri, 21 Nov 2003 06:19:03 -0400 (AST) Received: (qmail 31948 invoked by uid 0); 21 Nov 2003 10:18:59 -0000 Received: from unknown (HELO siaco.id.pl) (213.25.114.8) by smtp.id.pl with SMTP; 21 Nov 2003 10:18:58 -0000 Received: (qmail 1823 invoked by uid 1000); 21 Nov 2003 10:18:54 -0000 Date: Fri, 21 Nov 2003 11:18:54 +0100 From: Ryszard Lach To: Tom Lane Cc: pgsql-performance@postgresql.org, Bruce Momjian Subject: Re: duration logging setting in 7.4 Message-ID: <20031121101854.GA1106@siaco.id.pl> Reply-To: Ryszard Lach References: <20031117222058.GD3248@siaco.id.pl> <200311180237.hAI2b7s29417@candle.pha.pa.us> <20031118091646.GA8223@siaco.id.pl> <7948.1069373821@sss.pgh.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-2 Content-Disposition: inline In-Reply-To: <7948.1069373821@sss.pgh.pa.us> X-My-GPG-Key: echo | mail -s "send key pub" ryszard@lach.name User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/217 X-Sequence-Number: 4777 Hi, again. I've turned on only log_connections and log_statement. See the following log fragment (I've included lines only related to opening of new connection); Nov 21 11:06:44 postgres[3359]: [3-1] LOG: connection received: host= port= Nov 21 11:06:44 postgres[3359]: [4-1] LOG: connection authorized: user=pracuj database=pracuj Nov 21 11:06:44 postgres[3359]: [5-1] LOG: statement: set datestyle to 'ISO'; select version(), case when pg_encoding_to_char(1) = 'SQL_ASCII' then 'UNKNOWN' else Nov 21 11:06:44 postgres[3359]: [5-2] getdatabaseencoding() end; Nov 21 11:06:44 postgres[3359]: [6-1] LOG: statement: Nov 21 11:06:44 postgres[3359]: [7-1] LOG: statement: select * from ... Nov 21 11:06:45 postgres[3376]: [3-1] LOG: connection received: host= port= Nov 21 11:06:45 postgres[3376]: [4-1] LOG: connection authorized: user=pracuj database=pracuj Nov 21 11:06:45 postgres[3376]: [5-1] LOG: statement: set datestyle to 'ISO'; select version(), case when pg_encoding_to_char(1) = 'S QL_ASCII' then ' else Nov 21 11:06:45 postgres[3376]: [5-2] getdatabaseencoding() end; Nov 21 11:06:45 postgres[3376]: [6-1] LOG: statement: It seems, that empty statements are generated during opening of connection. Please, note also: 1. We are using an older jdbc driver (pgjdbc2) 2. We ar using encoding in URL (jdbc:postgresql://localhost:5432/database?charSet=iso-8859-1) Richard. -- "First they ignore you. Then they laugh at you. Then they fight you. Then you win." - Mohandas Gandhi. From pgsql-performance-owner@postgresql.org Fri Nov 21 10:15:00 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CD832D1D790; Fri, 21 Nov 2003 14:14:54 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 90152-01; Fri, 21 Nov 2003 10:14:27 -0400 (AST) Received: from bramble.mmrd.com (unknown [65.217.53.66]) by svr1.postgresql.org (Postfix) with ESMTP id 80337D1D7BD; Fri, 21 Nov 2003 10:14:23 -0400 (AST) Received: from thorn.mmrd.com (thorn.mmrd.com [172.25.10.100]) by bramble.mmrd.com (8.12.8/8.12.8) with ESMTP id hALDL7cM007075; Fri, 21 Nov 2003 08:21:07 -0500 Received: from gnvex001.mmrd.com (gnvex001.mmrd.com [192.168.3.55]) by thorn.mmrd.com (8.11.6/8.11.6) with ESMTP id hALEEEl07166; Fri, 21 Nov 2003 09:14:15 -0500 Received: from camel.mmrd.com ([172.25.5.213]) by gnvex001.mmrd.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2657.72) id V3JR4X5P; Fri, 21 Nov 2003 09:14:13 -0500 Subject: Re: [HACKERS] More detail on settings for pgavd? From: Robert Treat To: "Matthew T. O'Connor" Cc: Josh Berkus , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development In-Reply-To: <3FBD5EEF.7060205@zeut.net> References: <200311181558.45847.josh@agliodbs.com> <3FBC6BED.9090809@myrealbox.com> <3FBCD013.3030402@zeut.net> <200311200918.30838.josh@agliodbs.com> <3FBD5EEF.7060205@zeut.net> Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.8 Date: 21 Nov 2003 09:14:14 -0500 Message-Id: <1069424054.29672.12462.camel@camel> Mime-Version: 1.0 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/186 X-Sequence-Number: 4746 On Thu, 2003-11-20 at 19:40, Matthew T. O'Connor wrote: > I'm open to discussion on changing the defaults. Perhaps what it would > be better to use some non-linear (perhaps logorithmic) scaling factor. > So that you wound up with something roughly like this: > > #tuples activity% for vacuum > 1k 100% > 10k 70% > 100k 45% > 1M 20% > 10M 10% > 100M 8% > Just thinking out loud here, so disregard if you think its chaff but... if we had a system table pg_avd_defaults that held what we generally consider the best default percentages based on reltuples/pages, and added a column to pg_class (could be some place better but..) which could hold an overriding percentage, you could then have a column added to pg_stat_all_tables called vacuum_percentage, which would be a coalesce of the override percentage or the default percentages based on rel_tuples (or rel_pages). This would give autovacuum a place to look for each table as to when it should vacuum, and gives administrators the option to tweak it on a per table basis if they find they need a specific table to vacuum at a different rate than the "standard". Robert Treat -- Build A Brighter Lamp :: Linux Apache {middleware} PostgreSQL From pgsql-performance-owner@postgresql.org Mon Nov 24 14:53:16 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1A60CD1CC81; Fri, 21 Nov 2003 14:32:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 91583-02; Fri, 21 Nov 2003 10:31:52 -0400 (AST) Received: from mta5.srv.hcvlny.cv.net (mta5.srv.hcvlny.cv.net [167.206.5.78]) by svr1.postgresql.org (Postfix) with ESMTP id 74552D1D221; Fri, 21 Nov 2003 10:31:48 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta5.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HOP00L0DHP29B@mta5.srv.hcvlny.cv.net>; Fri, 21 Nov 2003 09:31:52 -0500 (EST) Date: Fri, 21 Nov 2003 09:31:49 -0500 From: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: <1069424054.29672.12462.camel@camel> To: Robert Treat Cc: Josh Berkus , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBE21D5.20508@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <200311181558.45847.josh@agliodbs.com> <3FBC6BED.9090809@myrealbox.com> <3FBCD013.3030402@zeut.net> <200311200918.30838.josh@agliodbs.com> <3FBD5EEF.7060205@zeut.net> <1069424054.29672.12462.camel@camel> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/216 X-Sequence-Number: 4776 Robert Treat wrote: >Just thinking out loud here, so disregard if you think its chaff but... >if we had a system table pg_avd_defaults > [snip] As long as pg_autovacuum remains a contrib module, I don't think any changes to the system catelogs will be make. If pg_autovacuum is deemed ready to move out of contrib, then we can talk about the above. From pgsql-performance-owner@postgresql.org Fri Nov 21 10:55:35 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2C4E9D1B520 for ; Fri, 21 Nov 2003 14:55:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 92633-03 for ; Fri, 21 Nov 2003 10:55:05 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id AAC17D1CC81 for ; Fri, 21 Nov 2003 10:55:00 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hALEsW19012145; Fri, 21 Nov 2003 09:54:33 -0500 (EST) To: Ryszard Lach Cc: pgsql-performance@postgresql.org, Bruce Momjian Subject: Re: duration logging setting in 7.4 In-reply-to: <20031121101854.GA1106@siaco.id.pl> References: <20031117222058.GD3248@siaco.id.pl> <200311180237.hAI2b7s29417@candle.pha.pa.us> <20031118091646.GA8223@siaco.id.pl> <7948.1069373821@sss.pgh.pa.us> <20031121101854.GA1106@siaco.id.pl> Comments: In-reply-to Ryszard Lach message dated "Fri, 21 Nov 2003 11:18:54 +0100" Date: Fri, 21 Nov 2003 09:54:32 -0500 Message-ID: <12144.1069426472@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/187 X-Sequence-Number: 4747 Ryszard Lach writes: > It seems, that empty statements are generated during opening of > connection. Hmm. Try asking about that on the pgsql-jdbc list. I think the JDBC driver must actually be sending empty commands. Looking at the backend code, I realize that 7.4 will emit LOG: entries for empty query strings, whereas prior releases would not. This isn't a bug IMHO, but it does explain why you are noticing output that wasn't there before you updated to 7.4. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Nov 21 10:56:54 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 20048D1B520; Fri, 21 Nov 2003 14:56:49 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 92464-05; Fri, 21 Nov 2003 10:56:21 -0400 (AST) Received: from mta10.srv.hcvlny.cv.net (mta10.srv.hcvlny.cv.net [167.206.5.85]) by svr1.postgresql.org (Postfix) with ESMTP id ED9BCD1D7C7; Fri, 21 Nov 2003 10:56:16 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta10.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HOP0002IITUPF@mta10.srv.hcvlny.cv.net>; Fri, 21 Nov 2003 09:56:21 -0500 (EST) Date: Fri, 21 Nov 2003 09:56:17 -0500 From: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: <200311202224.45065.josh@agliodbs.com> To: Josh Berkus Cc: Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBE2791.8040404@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <200311181558.45847.josh@agliodbs.com> <200311200918.30838.josh@agliodbs.com> <3FBD5EEF.7060205@zeut.net> <200311202224.45065.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/188 X-Sequence-Number: 4748 Josh Berkus wrote: >Matthew, > > >>True, but I think it would be one hour once, rather than 30 minutes 4 >>times. >> >> >Well, generally it would be about 6-8 times at 2-4 minutes each. > > Are you saying that you can vacuum a 1 million row table in 2-4 minutes? While a vacuum of the same table with an additional 1 million dead tuples would take an hour? >>This is one of the things I had hoped to add to pg_autovacuum, but never >>got to. In addition to just the information from the stats collector on >>inserts updates and deletes, pg_autovacuum should also look at the FSM, >>and make decisions based on it. Anyone looking for a project? >> >> >Hmmm ... I think that's the wrong approach. Once your database is populated, >it's very easy to determine how to set the FSM for a given pg_avd level. If >you're vacuuming after 20% updates, for example, just set fsm_pages to 20% of >the total database pages plus growth & safety margins. > > Ok. >I'd be really reluctant to base pv-avd frequency on the fsm settings instead. >What if the user loads 8GB of data but leaves fsm_pages at the default of >10,000? You can't do much with that; you'd have to vacuum if even 1% of the >data changed. > Ok, but as you said above it's very easy to set the FSM once you know your db size. >The other problem is that calculating data pages from a count of >updates+deletes would require pg_avd to keep more statistics and do more math >for every table. Do we want to do this? > > I would think the math is simple enough to not be a big problem. Also, I did not recommend looking blindly at the FSM as our guide, rather consulting it as another source of information as to when it would be useful to vacuum. I don't have a good plan as to how to incorporate this data, but to a large extent the FSM already tracks table activity and gives us the most accurate answer about storage growth (short of using something like contrib/pgstattuple which takes nearly the same amount of time as an actual vacuum) >>But I can't imagine that 2% makes any difference on a large table. In >>fact I would think that 10-15% would hardly be noticable, beyond that >>I'm not sure. >> >> >I've seen performance lag at 10% of records, especially in tables where both >update and select activity focus on one subset of the table (calendar tables, >for example). > > Ok. >>Valid points, and again I think this points to the fact that >>pg_autovacuum needs to be more configurable. Being able to set >>different thresholds for different tables will help considerably. In >>fact, you may find that some tables should have a vac threshold much >>larger than the analyze thresold, while other tables might want the >>opposite. >> >> >Sure. Though I think we can make the present configuration work with a little >adjustment of the numbers. I'll have a chance to test on production >databases soon. > > I look forward to hearing results from your testing. >>I would be surprized if you can notice the difference between a vacuum >>analyze and a vacuum, especially on large tables. >> >> >It's substantial for tables with high statistics settings. A 1,000,000 row >table with 5 columns set to statistics=250 can take 3 minutes to analyze on a >medium-grade server. > > In my testing, I never changed the default statistics settings. >>I think you need two separate schedules. There are lots of times where >>a vacuum doesn't help, and an analyze is all that is needed >> >> >Agreed. And I've just talked to a client who may want to use pg_avd's ANALYZE >scheduling but not use vacuum at all. BTW, I think we should have a setting >for this; for example, if -V is -1, don't vacuum. > > That would be nice. Easy to add, and something I never thought of.... >>I'm open to discussion on changing the defaults. Perhaps what it would >>be better to use some non-linear (perhaps logorithmic) scaling factor. >> >> >That would be cool, too. Though a count of data pages would be a better >scale than a count of rows, and equally obtainable from pg_class. > > But we track tuples because we can compare against the count given by the stats system. I don't know of a way (other than looking at the FSM, or contrib/pgstattuple ) to see how many dead pages exist. From pgsql-performance-owner@postgresql.org Fri Nov 21 11:07:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5A138D1B520; Fri, 21 Nov 2003 15:07:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 92477-07; Fri, 21 Nov 2003 11:07:13 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id 97813D1D7BD; Fri, 21 Nov 2003 11:07:08 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Fri, 21 Nov 2003 08:07:04 -0700 Message-ID: <3FBE2A0E.3010308@myrealbox.com> Date: Fri, 21 Nov 2003 20:36:54 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "Matthew T. O'Connor" Cc: Josh Berkus , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development Subject: Re: [HACKERS] More detail on settings for pgavd? References: <200311181558.45847.josh@agliodbs.com> <200311200918.30838.josh@agliodbs.com> <3FBD5EEF.7060205@zeut.net> <200311202224.45065.josh@agliodbs.com> <3FBE2791.8040404@zeut.net> In-Reply-To: <3FBE2791.8040404@zeut.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/189 X-Sequence-Number: 4749 Matthew T. O'Connor wrote: > But we track tuples because we can compare against the count given by > the stats system. I don't know of a way (other than looking at the FSM, > or contrib/pgstattuple ) to see how many dead pages exist. I think making pg_autovacuum dependent of pgstattuple is very good idea. Probably it might be a good idea to extend pgstattuple to return pages that are excessively contaminated and clean them ASAP. Step by step getting closer to daemonized vacuum. Shridhar From pgsql-performance-owner@postgresql.org Fri Nov 21 11:18:16 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5453FD1D767; Fri, 21 Nov 2003 15:18:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 94367-04; Fri, 21 Nov 2003 11:17:33 -0400 (AST) Received: from mta1.srv.hcvlny.cv.net (mta1.srv.hcvlny.cv.net [167.206.5.67]) by svr1.postgresql.org (Postfix) with ESMTP id 4F50BD1D450; Fri, 21 Nov 2003 11:17:29 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta1.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HOP00KH2JTPEQ@mta1.srv.hcvlny.cv.net>; Fri, 21 Nov 2003 10:17:50 -0500 (EST) Date: Fri, 21 Nov 2003 10:17:31 -0500 From: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: <3FBE2A0E.3010308@myrealbox.com> To: Shridhar Daithankar Cc: Josh Berkus , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBE2C8B.9040304@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <200311181558.45847.josh@agliodbs.com> <200311200918.30838.josh@agliodbs.com> <3FBD5EEF.7060205@zeut.net> <200311202224.45065.josh@agliodbs.com> <3FBE2791.8040404@zeut.net> <3FBE2A0E.3010308@myrealbox.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/190 X-Sequence-Number: 4750 Shridhar Daithankar wrote: > Matthew T. O'Connor wrote: > >> But we track tuples because we can compare against the count given by >> the stats system. I don't know of a way (other than looking at the >> FSM, or contrib/pgstattuple ) to see how many dead pages exist. > > I think making pg_autovacuum dependent of pgstattuple is very good idea. Only if pgstattuple can become much cheaper than it is now. Based on the testing I did when I wrote pg_autovacuum, pgstattuple cost nearly the same amount as a regular vacuum. Given that, what have we gained from that work? Wouldn't it just be better to run a vacuum and actually reclaim space rather than running pgstattuple, and just look and see if there is free space to be reclaimed? Perhaps we could use pgstattuple ocasionally to see if we are going a good job of keeping the amount of dead space to a reasonable level, but I'm still not really sure about this. > Probably it might be a good idea to extend pgstattuple to return pages > that are excessively contaminated and clean them ASAP. Step by step > getting closer to daemonized vacuum. I don't know of anyway to clean a particular set of pages. This is something that has been talked about (partial vacuums and such), but I think Tom has raised several issues with it, I don't remember the details. Right now the only tool we have to reclaim space is vacuum, a whole table at a time. From pgsql-performance-owner@postgresql.org Fri Nov 21 13:10:13 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id DB40AD1B555; Fri, 21 Nov 2003 17:10:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 42743-08; Fri, 21 Nov 2003 13:09:37 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 3C18BD1B54E; Fri, 21 Nov 2003 13:09:36 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3955426; Fri, 21 Nov 2003 09:10:21 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: "Matthew T. O'Connor" , Robert Treat Subject: Re: [HACKERS] More detail on settings for pgavd? Date: Fri, 21 Nov 2003 09:09:00 -0800 User-Agent: KMail/1.4.3 Cc: Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development References: <200311181558.45847.josh@agliodbs.com> <1069424054.29672.12462.camel@camel> <3FBE21D5.20508@zeut.net> In-Reply-To: <3FBE21D5.20508@zeut.net> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311210909.00978.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/191 X-Sequence-Number: 4751 Matthew, > As long as pg_autovacuum remains a contrib module, I don't think any > changes to the system catelogs will be make. If pg_autovacuum is > deemed ready to move out of contrib, then we can talk about the above. But we could create a config file that would store stuff in a flatfile table, OR we could add our own "system table" that would be created when one "initializes" pg_avd. Just an idea. Mind you, I'm not so sure that we want to focus immediately on per-table settings. I think that we want to get the "automatic" settings working fairly well first; a lot of new DBAs would use the per-table settings to shoot themselves in the foot. So we need to be able to make a strong recommendation to "try the automatic settings first." > Are you saying that you can vacuum a 1 million row table in 2-4 > minutes? While a vacuum of the same table with an additional 1 million > dead tuples would take an hour? I'm probably exaggerating. I do know that I can vacuum a fairly clean 1-5 million row table in less than 4 mintues. I've never let such a table get to 50% dead tuples, so I don't really know how long that takes. Call me a coward if you like ... > >I'd be really reluctant to base pv-avd frequency on the fsm settings > > instead. What if the user loads 8GB of data but leaves fsm_pages at the > > default of 10,000? You can't do much with that; you'd have to vacuum if > > even 1% of the data changed. > > Ok, but as you said above it's very easy to set the FSM once you know > your db size. Actually, thinking about this I realize that PG_AVD and the Perl-based postgresql.conf configuration script I was working on (darn, who was doing that with me?) need to go togther. With pg_avd, setting max_fsm_pages is very easy; without it its a bit of guesswork. So I think we can do this: for 'auto' settings: If max_fsm_pages is between 13% and 100% of the total database pages, then set the vacuum scale factor to match 3/4 of the fsm_pages setting, e.g. database = 18,000,000 data pages; max_fsm_pages = 3,600,000; set vacuum scale factor = 3.6mil/18mil * 3/4 = 0.15 If max_fsm_pages is less than 13% of database pages, issue a warning to the user (log it, if possible) and set scale factor to 0.1. If it's greater than 100% set it to 1 and leave it alone. > I don't have a good plan as to how to incorporate > this data, but to a large extent the FSM already tracks table activity > and gives us the most accurate answer about storage growth (short of > using something like contrib/pgstattuple which takes nearly the same > amount of time as an actual vacuum) I don't really think we need to do dynamic monitoring at this point. It would be a lot of engineering to check data page pollution without having significant performance impact. It's doable, but something I think we should hold off until version 3. It would mean hacking the FSM, which is a little beyond me right now. > In my testing, I never changed the default statistics settings. Ah. Well, a lot of users do to resolve query problems. > But we track tuples because we can compare against the count given by > the stats system. I don't know of a way (other than looking at the FSM, > or contrib/pgstattuple ) to see how many dead pages exist. No, but for scaling you don't need the dynamic count of tuples or of dead tuples; pg_class holds a reasonable accurate count of pages per table as of last vacuum. -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Fri Nov 21 17:34:19 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D294ED1D660; Fri, 21 Nov 2003 21:34:17 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 90740-10; Fri, 21 Nov 2003 17:33:48 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 04671D1CA7F; Fri, 21 Nov 2003 17:33:46 -0400 (AST) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3956660; Fri, 21 Nov 2003 13:34:19 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? Date: Fri, 21 Nov 2003 13:23:18 -0800 User-Agent: KMail/1.4.3 Cc: Robert Treat , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development References: <200311181558.45847.josh@agliodbs.com> <200311210909.00978.josh@agliodbs.com> <3FBE8289.1040009@zeut.net> In-Reply-To: <3FBE8289.1040009@zeut.net> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200311211323.19003.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/193 X-Sequence-Number: 4753 Matthew, > Actually, this might be a necessary addition as pg_autovacuum currently= =20 > suffers from the startup transients that the FSM used to suffer from,=20 > that is, it doesn't remember anything that happened the last time it=20 > ran. A pg_autovacuum database could also be used to store thresholds=20 > and counts from the last time it ran. I don't see how a seperate database is better than a table in the databases= .,=20 except that it means scanning only one table and not one per database. Fo= r=20 one thing, making it a seperate database could make it hard to back up and= =20 move your database+pg_avd config. But I don't feel strongly about it. > Where are you getting 13% from?=20 13% * 3/4 ~~ 10% And I think both of use agree that vacuuming tables with less than 10% chan= ges=20 is excessive and could lead to problems on its own, like overlapping vacuum= s. > Do you know of an easy way to get a=20 > count of the total pages used by a whole cluster? Select sum(relpages) from pg_class. > I do like the concept though as long as we find good values for=20 > min_fsm_percentage and min_autovac_scaling_factor. See above. I propose 0.13 and 0.1 > Which we already keep a copy of inside of pg_autovacuum, and update=20 > after we issue a vacuum. Even easier then. BTW, do we have any provisions to avoid overlapping vacuums? That is, to= =20 prevent a second vacuum on a table if an earlier one is still running? --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Fri Nov 21 17:25:39 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6F0ECD1D1C6; Fri, 21 Nov 2003 21:25:35 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 90692-05; Fri, 21 Nov 2003 17:25:05 -0400 (AST) Received: from mta2.srv.hcvlny.cv.net (mta2.srv.hcvlny.cv.net [167.206.5.68]) by svr1.postgresql.org (Postfix) with ESMTP id 73CE2D1D49D; Fri, 21 Nov 2003 17:25:03 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta2.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HOQ0099V0TOFV@mta2.srv.hcvlny.cv.net>; Fri, 21 Nov 2003 16:25:03 -0500 (EST) Date: Fri, 21 Nov 2003 16:24:25 -0500 From: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: <200311210909.00978.josh@agliodbs.com> To: Josh Berkus Cc: Robert Treat , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBE8289.1040009@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <200311181558.45847.josh@agliodbs.com> <1069424054.29672.12462.camel@camel> <3FBE21D5.20508@zeut.net> <200311210909.00978.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/192 X-Sequence-Number: 4752 Josh Berkus wrote: >Matthew, > > >But we could create a config file that would store stuff in a flatfile table, >OR we could add our own "system table" that would be created when one >"initializes" pg_avd. > > I don't want to add tables to existing databases, as I consider that clutter and I never like using tools that clutter my production databases. I had considered using a pg_autovacuum database that if found, would store customized settings for individual tables / databases. Dunno if this is a good idea, but it might make a good stopgap until people are comfortable modifying the system catalogs for autovacuum. Actually, this might be a necessary addition as pg_autovacuum currently suffers from the startup transients that the FSM used to suffer from, that is, it doesn't remember anything that happened the last time it ran. A pg_autovacuum database could also be used to store thresholds and counts from the last time it ran. >Just an idea. Mind you, I'm not so sure that we want to focus immediately on >per-table settings. I think that we want to get the "automatic" settings >working fairly well first; a lot of new DBAs would use the per-table settings >to shoot themselves in the foot. So we need to be able to make a strong >recommendation to "try the automatic settings first." > > I agree in principle, question is what are the best settings, I still think it will be hard to find a one size fits all, but I'm sure we can do better than what we have. >Actually, thinking about this I realize that PG_AVD and the Perl-based >postgresql.conf configuration script I was working on (darn, who was doing >that with me?) need to go togther. With pg_avd, setting max_fsm_pages is >very easy; without it its a bit of guesswork. > >So I think we can do this: for 'auto' settings: > >If max_fsm_pages is between 13% and 100% of the total database pages, then set >the vacuum scale factor to match 3/4 of the fsm_pages setting, e.g. >database = 18,000,000 data pages; >max_fsm_pages = 3,600,000; >set vacuum scale factor = 3.6mil/18mil * 3/4 = 0.15 > > Where are you getting 13% from? Do you know of an easy way to get a count of the total pages used by a whole cluster? I guess we can just iterate over all the tables in all the databases and sum up the total num of pages. We already iterate over them all, we just don't sum it up. >If max_fsm_pages is less than 13% of database pages, issue a warning to the >user (log it, if possible) and set scale factor to 0.1. If it's greater >than 100% set it to 1 and leave it alone. > > Again I ask where 13% is coming from and also where is 0.1 coming from? I assume these are your best guesses right now, but not more than that. I do like the concept though as long as we find good values for min_fsm_percentage and min_autovac_scaling_factor. >>But we track tuples because we can compare against the count given by >>the stats system. I don't know of a way (other than looking at the FSM, >>or contrib/pgstattuple ) to see how many dead pages exist. >> >> >No, but for scaling you don't need the dynamic count of tuples or of dead >tuples; pg_class holds a reasonable accurate count of pages per table as of >last vacuum. > > Which we already keep a copy of inside of pg_autovacuum, and update after we issue a vacuum. From pgsql-performance-owner@postgresql.org Fri Nov 21 18:00:44 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1B0C1D1B53A; Fri, 21 Nov 2003 22:00:43 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 93371-03; Fri, 21 Nov 2003 18:00:13 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 6E64CD1B515; Fri, 21 Nov 2003 18:00:10 -0400 (AST) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3956811; Fri, 21 Nov 2003 14:00:59 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? Date: Fri, 21 Nov 2003 13:49:58 -0800 User-Agent: KMail/1.4.3 Cc: Robert Treat , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development References: <200311181558.45847.josh@agliodbs.com> <200311211323.19003.josh@agliodbs.com> <3FBE891D.8010309@zeut.net> In-Reply-To: <3FBE891D.8010309@zeut.net> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200311211349.58788.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/195 X-Sequence-Number: 4755 Matthew, > Basically, I don't like the idea of modifying users databases, besides,= =20 > in the long run most of what needs to be tracked will be moved to the=20 > system catalogs. I kind of consider the pg_autvacuum database to=20 > equivalent to the changes that will need to be made to the system catalog= s. OK. As I said, I don't feel strongly about it. > I certainly agree that less than 10% would be excessive, I still feel=20 > that 10% may not be high enough though. That's why I kinda liked the=20 > sliding scale I mentioned earlier, because I agree that for very large=20 > tables, something as low as 10% might be useful, but most tables in a=20 > database would not be that large. Yes, but I thought that we were taking care of that through the "threshold"= =20 value? A sliding scale would also be OK. However, that would definitely require = a=20 leap to storing per-table pg_avd statistics and settings. > Only that pg_autovacuum isn't smart enough to kick off more than one=20 > vacuum at a time. Basically, pg_autovacuum issues a vacuum on a table=20 > and waits for it to finish, then check the next table in it's list to=20 > see if it needs to be vacuumed, if so, it does it and waits for that=20 > vacuum to finish.=20 OK, then, we just need to detect the condition of the vacuums "piling up"= =20 because they are happening too often. --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Fri Nov 21 17:53:23 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 7E504D1D2BB; Fri, 21 Nov 2003 21:53:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 92181-07; Fri, 21 Nov 2003 17:52:50 -0400 (AST) Received: from mta1.srv.hcvlny.cv.net (mta1.srv.hcvlny.cv.net [167.206.5.67]) by svr1.postgresql.org (Postfix) with ESMTP id B6D5FD1D660; Fri, 21 Nov 2003 17:52:47 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta1.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HOQ00GGU24IBC@mta1.srv.hcvlny.cv.net>; Fri, 21 Nov 2003 16:53:08 -0500 (EST) Date: Fri, 21 Nov 2003 16:52:29 -0500 From: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: <200311211323.19003.josh@agliodbs.com> To: josh@agliodbs.com Cc: Robert Treat , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBE891D.8010309@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <200311181558.45847.josh@agliodbs.com> <200311210909.00978.josh@agliodbs.com> <3FBE8289.1040009@zeut.net> <200311211323.19003.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/194 X-Sequence-Number: 4754 Josh Berkus wrote: >Matthew, > > > >I don't see how a seperate database is better than a table in the databases., >except that it means scanning only one table and not one per database. For >one thing, making it a seperate database could make it hard to back up and >move your database+pg_avd config. > > Basically, I don't like the idea of modifying users databases, besides, in the long run most of what needs to be tracked will be moved to the system catalogs. I kind of consider the pg_autvacuum database to equivalent to the changes that will need to be made to the system catalogs. I guess it could make it harder to backup if you are moving your database between clusters. Perhaps, if you create a pg_autovacuum schema inside of your database then we would could use that. I just don't like tools that drop things into your database. >>Where are you getting 13% from? >> >> > >13% * 3/4 ~~ 10% > >And I think both of use agree that vacuuming tables with less than 10% changes >is excessive and could lead to problems on its own, like overlapping vacuums. > > > I certainly agree that less than 10% would be excessive, I still feel that 10% may not be high enough though. That's why I kinda liked the sliding scale I mentioned earlier, because I agree that for very large tables, something as low as 10% might be useful, but most tables in a database would not be that large. >> Do you know of an easy way to get a >>count of the total pages used by a whole cluster? >> >> > >Select sum(relpages) from pg_class. > > > duh.... >BTW, do we have any provisions to avoid overlapping vacuums? That is, to >prevent a second vacuum on a table if an earlier one is still running? > > > Only that pg_autovacuum isn't smart enough to kick off more than one vacuum at a time. Basically, pg_autovacuum issues a vacuum on a table and waits for it to finish, then check the next table in it's list to see if it needs to be vacuumed, if so, it does it and waits for that vacuum to finish. There was some discussion of issuing concurrent vacuum against different tables, but it was decided that since vacuum is I/O bound, it would only make sense to issue concurrent vacuums that were on different spindles, which is not something I wanted to get into. Also, given the recent talk about how vacuum is still such a performance hog, I can't imagine what multiple concurrent vacuums would do to performance. Maybe as 7.5 develops and many of the vacuum performance issues are addressed, we can revisit this question. From pgsql-performance-owner@postgresql.org Fri Nov 21 18:41:24 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2BD7ED1C4D4; Fri, 21 Nov 2003 22:41:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 99775-06; Fri, 21 Nov 2003 18:40:48 -0400 (AST) Received: from mta1.srv.hcvlny.cv.net (mta1.srv.hcvlny.cv.net [167.206.5.67]) by svr1.postgresql.org (Postfix) with ESMTP id E9ED6D1B553; Fri, 21 Nov 2003 18:40:45 -0400 (AST) Received: from zeut.net (ool-4352919e.dyn.optonline.net [67.82.145.158]) by mta1.srv.hcvlny.cv.net (iPlanet Messaging Server 5.2 HotFix 1.16 (built May 14 2003)) with ESMTP id <0HOQ00L974CHVV@mta1.srv.hcvlny.cv.net>; Fri, 21 Nov 2003 17:41:07 -0500 (EST) Date: Fri, 21 Nov 2003 17:40:45 -0500 From: "Matthew T. O'Connor" Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: <200311211349.58788.josh@agliodbs.com> To: josh@agliodbs.com Cc: Robert Treat , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development Message-id: <3FBE946D.4050708@zeut.net> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 References: <200311181558.45847.josh@agliodbs.com> <200311211323.19003.josh@agliodbs.com> <3FBE891D.8010309@zeut.net> <200311211349.58788.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/196 X-Sequence-Number: 4756 Josh Berkus wrote: >Matthew, > > >>I certainly agree that less than 10% would be excessive, I still feel >>that 10% may not be high enough though. That's why I kinda liked the >>sliding scale I mentioned earlier, because I agree that for very large >>tables, something as low as 10% might be useful, but most tables in a >>database would not be that large. >> >> > >Yes, but I thought that we were taking care of that through the "threshold" >value? > > Well the threshold is a combination of the base value and the scaling factor which you are proposing is 0.1, so the threshold is base + (scaling factor)(num of tuples) So with the default base of 1000 and your 0.1 you would have this: Num Rows threshold Percent 1,000 1,100 110% 10,000 2,000 20% 100,000 11,000 11% 1,000,000 102,000 10% I don't like how that looks, hence the thought of some non-linear scaling factor that would still allow the percent to reach 10%, but at a slower rate, perhaps just a larger base value would suffice, but I think small table performance is going to suffer much above 1000. Anyone else have an opinion on the table above? Good / Bad / Indifferent? >A sliding scale would also be OK. However, that would definitely require a >leap to storing per-table pg_avd statistics and settings. > > > I don't think it would, it would correlate the scaling factor with the number of tuples, no per-table settings required. >>Only that pg_autovacuum isn't smart enough to kick off more than one >>vacuum at a time. Basically, pg_autovacuum issues a vacuum on a table >>and waits for it to finish, then check the next table in it's list to >>see if it needs to be vacuumed, if so, it does it and waits for that >>vacuum to finish. >> >> > >OK, then, we just need to detect the condition of the vacuums "piling up" >because they are happening too often. > > > That would be good to look into at some point, especially if vacuum is going to get slower as a result of the page loop delay patch that has been floating around. From pgsql-performance-owner@postgresql.org Fri Nov 21 19:05:39 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 02784D1D524 for ; Fri, 21 Nov 2003 23:05:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 05523-01 for ; Fri, 21 Nov 2003 19:04:58 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 68AF7D1B53A for ; Fri, 21 Nov 2003 19:04:55 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hALN4u19015609; Fri, 21 Nov 2003 18:04:57 -0500 (EST) To: josh@agliodbs.com Cc: "Matthew T. O'Connor" , Robert Treat , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development Subject: Re: [HACKERS] More detail on settings for pgavd? In-reply-to: <200311211323.19003.josh@agliodbs.com> References: <200311181558.45847.josh@agliodbs.com> <200311210909.00978.josh@agliodbs.com> <3FBE8289.1040009@zeut.net> <200311211323.19003.josh@agliodbs.com> Comments: In-reply-to Josh Berkus message dated "Fri, 21 Nov 2003 13:23:18 -0800" Date: Fri, 21 Nov 2003 18:04:56 -0500 Message-ID: <15608.1069455896@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/197 X-Sequence-Number: 4757 Josh Berkus writes: > BTW, do we have any provisions to avoid overlapping vacuums? That is, to > prevent a second vacuum on a table if an earlier one is still running? Yes, VACUUM takes a lock that prevents another VACUUM on the same table. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Nov 21 19:54:14 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C30DED1D7BC; Fri, 21 Nov 2003 23:54:11 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 07596-06; Fri, 21 Nov 2003 19:53:43 -0400 (AST) Received: from arbor.net (division.aa.arbor.net [204.181.64.2]) by svr1.postgresql.org (Postfix) with ESMTP id 625C5D1D59A; Fri, 21 Nov 2003 19:53:40 -0400 (AST) Received: by arbor.net (Postfix, from userid 1065) id 14BE22A893; Fri, 21 Nov 2003 18:53:27 -0500 (EST) Received: from localhost (localhost [127.0.0.1]) by arbor.net (Postfix) with ESMTP id 0D5FC2A88C; Fri, 21 Nov 2003 18:53:27 -0500 (EST) Date: Fri, 21 Nov 2003 18:53:26 -0500 (EST) From: Chester Kustarz To: "Matthew T. O'Connor" Cc: pgsql-performance@postgresql.org, PostgreSQL-development Subject: Re: [HACKERS] More detail on settings for pgavd? In-Reply-To: <3FBE891D.8010309@zeut.net> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/198 X-Sequence-Number: 4758 On Fri, 21 Nov 2003, Matthew T. O'Connor wrote: > >> Do you know of an easy way to get a > >>count of the total pages used by a whole cluster? > > > >Select sum(relpages) from pg_class. You might want to exclude indexes from this calculation. Some large read only tables might have indexes larger than the tables themselves. From pgsql-hackers-owner@postgresql.org Fri Nov 21 20:52:03 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 07EE6D1D809 for ; Sat, 22 Nov 2003 00:52:02 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 10327-08 for ; Fri, 21 Nov 2003 20:51:36 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id E2D6ED1D7C5 for ; Fri, 21 Nov 2003 20:51:32 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id C055336E06; Fri, 21 Nov 2003 19:51:17 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1ANLzZ-0006LP-00; Fri, 21 Nov 2003 19:51:17 -0500 To: pgsql-hackers@postgresql.org Subject: Re: [PERFORM] More detail on settings for pgavd? References: <200311181558.45847.josh@agliodbs.com> <200311210909.00978.josh@agliodbs.com> <3FBE8289.1040009@zeut.net> <200311211323.19003.josh@agliodbs.com> <15608.1069455896@sss.pgh.pa.us> In-Reply-To: <15608.1069455896@sss.pgh.pa.us> From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 21 Nov 2003 19:51:17 -0500 Message-ID: <877k1t2pkq.fsf@stark.dyndns.tv> Lines: 17 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/1213 X-Sequence-Number: 47501 Tom Lane writes: > Josh Berkus writes: > > BTW, do we have any provisions to avoid overlapping vacuums? That is, to > > prevent a second vacuum on a table if an earlier one is still running? > > Yes, VACUUM takes a lock that prevents another VACUUM on the same table. The second vacuum waits for the lock to become available. If the situation got really bad there could end up being a growing queue of vacuums waiting. I'm not sure how likely this is as the subsequent vacuums appear to finish quite quickly though. But then the largest table I have to play with fits entirely in memory. -- greg From pgsql-performance-owner@postgresql.org Sat Nov 22 07:57:35 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E94AFD1CA8E for ; Sat, 22 Nov 2003 11:57:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 17885-08 for ; Sat, 22 Nov 2003 07:57:06 -0400 (AST) Received: from easily.co.uk (unknown [213.161.76.90]) by svr1.postgresql.org (Postfix) with ESMTP id 46C73D1D5D9 for ; Sat, 22 Nov 2003 07:57:02 -0400 (AST) Received: from [80.235.141.90] (account f4vo5dsy5djd HELO chuckie.co.uk) by easily.co.uk (CommuniGate Pro SMTP 4.1.3) with ESMTP id 35518619; Sat, 22 Nov 2003 11:56:34 +0000 Message-ID: <3FBF4EF3.6090507@chuckie.co.uk> Date: Sat, 22 Nov 2003 11:56:35 +0000 From: Nick Barr User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5b) Gecko/20030901 Thunderbird/0.2 X-Accept-Language: en-us, en MIME-Version: 1.0 To: josh@agliodbs.com, pgsql-performance@postgresql.org Subject: pg_autoconfig.pl Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/199 X-Sequence-Number: 4759 Heya, Sorry for no updates on the old pg_autoconfig script thing, been rather busy at work the last week or two :-( but i suppose it pays the bills. I do have a few days off work now so I can spend some time on finishing the first version off. The latest version can be found at http://www.chuckie.co.uk/postgresql/pg_autoconfig.txt Josh have you managed to put together the rest of the calculations yet? Thanks, Nick From pgsql-performance-owner@postgresql.org Mon Nov 24 14:57:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6F9C4D1D453; Sat, 22 Nov 2003 21:36:17 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 07852-03; Sat, 22 Nov 2003 17:35:48 -0400 (AST) Received: from sunsite.dcc.uchile.cl (sunsite.dcc.uchile.cl [192.80.24.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1273ED1D5FD; Sat, 22 Nov 2003 17:35:45 -0400 (AST) Received: from root@anakena.dcc.uchile.cl [192.80.24.3] by sunsite.dcc.uchile.cl (8.8.5/Main-DCCV8-Jo6) id SAA09191; Sat, 22 Nov 2003 18:35:32 -0300 (CDT) Received: from alvherre@localhost by anakena.dcc.uchile.cl (8.8.5/MainSec-DCCV8-Jo5) id LAA07875; Sat, 22 Nov 2003 11:46:37 -0300 (CLST) Date: Sat, 22 Nov 2003 11:46:37 -0300 From: Alvaro Herrera Munoz To: "Matthew T. O'Connor" Cc: Josh Berkus , Robert Treat , Shridhar Daithankar , pgsql-performance@postgresql.org, PostgreSQL-development Subject: Re: [HACKERS] More detail on settings for pgavd? Message-ID: <20031122144637.GC6016@dcc.uchile.cl> References: <200311181558.45847.josh@agliodbs.com> <1069424054.29672.12462.camel@camel> <3FBE21D5.20508@zeut.net> <200311210909.00978.josh@agliodbs.com> <3FBE8289.1040009@zeut.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3FBE8289.1040009@zeut.net> User-Agent: Mutt/1.4.1i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/222 X-Sequence-Number: 4782 On Fri, Nov 21, 2003 at 04:24:25PM -0500, Matthew T. O'Connor wrote: > I don't want to add tables to existing databases, as I consider that > clutter and I never like using tools that clutter my production > databases. [...] > > Actually, this might be a necessary addition as pg_autovacuum currently > suffers from the startup transients that the FSM used to suffer from, > that is, it doesn't remember anything that happened the last time it > ran. A pg_autovacuum database could also be used to store thresholds > and counts from the last time it ran. You could use the same approach the FSM uses: keep a file with the data, PGDATA/base/global/pg_fsm.cache. You don't need the data to be in a table after all ... -- Alvaro Herrera () Essentially, you're proposing Kevlar shoes as a solution for the problem that you want to walk around carrying a loaded gun aimed at your foot. (Tom Lane) From pgsql-performance-owner@postgresql.org Sat Nov 22 13:20:45 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1FBFAD1C96C for ; Sat, 22 Nov 2003 17:20:44 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 75938-06 for ; Sat, 22 Nov 2003 13:20:13 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id C79B2D1D630 for ; Sat, 22 Nov 2003 13:20:11 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3960326; Sat, 22 Nov 2003 09:20:51 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: Nick Barr , pgsql-performance@postgresql.org Subject: Re: pg_autoconfig.pl Date: Sat, 22 Nov 2003 09:19:23 -0800 User-Agent: KMail/1.4.3 References: <3FBF4EF3.6090507@chuckie.co.uk> In-Reply-To: <3FBF4EF3.6090507@chuckie.co.uk> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311220919.23695.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/200 X-Sequence-Number: 4760 Nick, > Josh have you managed to put together the rest of the calculations yet? Given that I spent most of November working on the PR for the 7.4 release, I've just started to think about it. As you can see, I'm thinking about dovetailing pg_avd and pg_autoconf. The difficult thing is to figure out settings for "bad" hardware setups. Like a 5GB database on a PIII500 + 256mb running 4 other pieces of major software (I've acctually seen this). Or a 500/minute OLTP database on a machine with 1 fixed disk. All the variables are going into a hash array, right? -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Sun Nov 23 23:51:39 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D2F49D1B531 for ; Mon, 24 Nov 2003 03:51:30 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 55336-10 for ; Sun, 23 Nov 2003 23:51:02 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id BFB02D1B4E1 for ; Sun, 23 Nov 2003 23:50:58 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hAO3ovNu052862 for ; Mon, 24 Nov 2003 03:50:57 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hAO3mBR1052409 for pgsql-performance@postgresql.org; Mon, 24 Nov 2003 03:48:11 GMT From: William Yu X-Newsgroups: comp.databases.postgresql.performance Subject: Maximum Possible Insert Performance? Date: Sun, 23 Nov 2003 19:48:13 -0800 Organization: Hub.Org Networking Services Lines: 12 Message-ID: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Complaints-To: usenet@news.hub.org User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en-us, en To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/201 X-Sequence-Number: 4761 My situation is this. We have a semi-production server where we pre-process data and then upload the finished data to our production servers. We need the fastest possible write performance. Having the DB go corrupt due to power loss/OS crash is acceptable because we can always restore from last night and re-run everything that was done since then. I already have fsync off. Short of buying more hardware -- which I will probably do anyways once I figure out whether I need more CPU, memory or disk -- what else can I do to max out the speed? Operation mix is about 50% select, 40% insert, 10% update. From pgsql-performance-owner@postgresql.org Mon Nov 24 00:22:14 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A818CD1C943 for ; Mon, 24 Nov 2003 04:22:06 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 77687-03 for ; Mon, 24 Nov 2003 00:21:35 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 8AB26D1D607 for ; Mon, 24 Nov 2003 00:21:34 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAO4LX19002409; Sun, 23 Nov 2003 23:21:33 -0500 (EST) To: William Yu Cc: pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? In-reply-to: References: Comments: In-reply-to William Yu message dated "Sun, 23 Nov 2003 19:48:13 -0800" Date: Sun, 23 Nov 2003 23:21:33 -0500 Message-ID: <2408.1069647693@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/202 X-Sequence-Number: 4762 William Yu writes: > [ we don't care about data integrity ] > I already have fsync off. Short of buying more hardware -- which I will > probably do anyways once I figure out whether I need more CPU, memory or > disk -- what else can I do to max out the speed? Operation mix is about > 50% select, 40% insert, 10% update. Batch operations so you commit more than one insert per transaction. (With fsync off, this isn't such a killer consideration as it would be with fsync on, but the per-transaction overhead is still nontrivial.) Get rid of as many integrity constraints as you feel can reasonably be postponed to the final upload. FK checks are particularly painful. Eliminate indexes where possible. Also (I hate to say this, but...) you should consider using Some Other Database. "I don't care about data integrity, only speed" sounds like a good fit to MySQL ... regards, tom lane From pgsql-performance-owner@postgresql.org Mon Nov 24 00:30:32 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 53FA7D1C96C for ; Mon, 24 Nov 2003 04:30:30 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 70237-09 for ; Mon, 24 Nov 2003 00:29:59 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id BA25AD1C951 for ; Mon, 24 Nov 2003 00:29:58 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3966870; Sun, 23 Nov 2003 20:30:44 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: William Yu , pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? Date: Sun, 23 Nov 2003 20:29:04 -0800 User-Agent: KMail/1.4.3 References: In-Reply-To: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311232029.04120.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/203 X-Sequence-Number: 4763 William, > I already have fsync off. Short of buying more hardware -- which I will > probably do anyways once I figure out whether I need more CPU, memory or > disk -- what else can I do to max out the speed? Operation mix is about > 50% select, 40% insert, 10% update. Disk. Multi-channel RAID is where it's at, and/or RAID with a great write cache enabled. For really fast updates, I'd suggest 6-disk or even 8-disk RAID 1+0. As soon as you have gobs of extra disk space, jack your checkpoint_buffers way up, like a couple of gigs. -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Mon Nov 24 00:42:38 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5A80ED1B8B9 for ; Mon, 24 Nov 2003 04:42:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 81247-04 for ; Mon, 24 Nov 2003 00:42:02 -0400 (AST) Received: from joeconway.com (66-146-172-86.skyriver.net [66.146.172.86]) by svr1.postgresql.org (Postfix) with ESMTP id E3DE3D1CAD1 for ; Mon, 24 Nov 2003 00:42:01 -0400 (AST) Received: from [192.168.5.3] (account jconway HELO joeconway.com) by joeconway.com (CommuniGate Pro SMTP 4.1.8) with ESMTP-TLS id 610270; Sun, 23 Nov 2003 20:39:54 -0800 Message-ID: <3FC18BAF.7070506@joeconway.com> Date: Sun, 23 Nov 2003 20:40:15 -0800 From: Joe Conway User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.3) Gecko/20030312 X-Accept-Language: en-us, en MIME-Version: 1.0 To: William Yu Cc: pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? References: In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/204 X-Sequence-Number: 4764 William Yu wrote: > My situation is this. We have a semi-production server where we > pre-process data and then upload the finished data to our production > servers. We need the fastest possible write performance. Having the DB > go corrupt due to power loss/OS crash is acceptable because we can > always restore from last night and re-run everything that was done since > then. If you can, use COPY -- it is far faster than INSERT. See: http://www.postgresql.org/docs/current/static/sql-copy.html HTH, Joe From pgsql-performance-owner@postgresql.org Mon Nov 24 01:51:20 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 09F61D1C951 for ; Mon, 24 Nov 2003 05:51:14 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 94076-02 for ; Mon, 24 Nov 2003 01:50:43 -0400 (AST) Received: from smtp-send.myrealbox.com (smtp-send.myrealbox.com [192.108.102.143]) by svr1.postgresql.org (Postfix) with ESMTP id C3A69D1D7C7 for ; Mon, 24 Nov 2003 01:50:41 -0400 (AST) Received: from myrealbox.com shridhar_daithankar@smtp-send.myrealbox.com [202.54.11.72] by smtp-send.myrealbox.com with NetMail SMTP Agent $Revision: 3.44 $ on Novell NetWare via secured & encrypted transport (TLS); Sun, 23 Nov 2003 22:50:41 -0700 Message-ID: <3FC19C2C.4050704@myrealbox.com> Date: Mon, 24 Nov 2003 11:20:36 +0530 From: Shridhar Daithankar User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 Cc: pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? References: In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/205 X-Sequence-Number: 4765 William Yu wrote: > My situation is this. We have a semi-production server where we > pre-process data and then upload the finished data to our production > servers. We need the fastest possible write performance. Having the DB > go corrupt due to power loss/OS crash is acceptable because we can > always restore from last night and re-run everything that was done since > then. > > I already have fsync off. Short of buying more hardware -- which I will > probably do anyways once I figure out whether I need more CPU, memory or > disk -- what else can I do to max out the speed? Operation mix is about > 50% select, 40% insert, 10% update. Mount WAL on RAM disk. WAL is most often hit area for heavy updates/inserts. If you spped that up, things should be pretty faster. A non-tried advice though. Given that you can afford a crash, I would say it is worth a try.. Shridhar From pgsql-performance-owner@postgresql.org Mon Nov 24 07:26:25 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0421AD1D989 for ; Mon, 24 Nov 2003 11:26:22 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 43333-09 for ; Mon, 24 Nov 2003 07:25:52 -0400 (AST) Received: from easily.co.uk (unknown [213.161.76.90]) by svr1.postgresql.org (Postfix) with ESMTP id F270CD1D8DA for ; Mon, 24 Nov 2003 07:25:48 -0400 (AST) Received: from [80.3.250.23] (HELO hare) by easily.co.uk (CommuniGate Pro SMTP 4.1.3) with SMTP id 35717045; Mon, 24 Nov 2003 11:25:31 +0000 Message-ID: <000c01c3b27d$91b4fec0$0300a8c0@hare> From: "Nick Barr" To: "Josh Berkus" , References: <3FBF4EF3.6090507@chuckie.co.uk> <200311220919.23695.josh@agliodbs.com> Subject: Re: pg_autoconfig.pl Date: Mon, 24 Nov 2003 11:24:47 -0000 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/206 X-Sequence-Number: 4766 ----- Original Message ----- From: "Josh Berkus" To: "Nick Barr" ; Sent: Saturday, November 22, 2003 5:19 PM Subject: Re: pg_autoconfig.pl > Nick, > > > Josh have you managed to put together the rest of the calculations yet? > > Given that I spent most of November working on the PR for the 7.4 release, > I've just started to think about it. As you can see, I'm thinking about > dovetailing pg_avd and pg_autoconf. > > The difficult thing is to figure out settings for "bad" hardware setups. Like > a 5GB database on a PIII500 + 256mb running 4 other pieces of major software > (I've acctually seen this). Or a 500/minute OLTP database on a machine > with 1 fixed disk. > > All the variables are going into a hash array, right? Yep that rights.$sys is the array name. And there are loads of different keys for the hash. I will try and get a listing out of the script sometime. Nick From pgsql-performance-owner@postgresql.org Mon Nov 24 12:03:35 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B7ACED1DA3B for ; Mon, 24 Nov 2003 13:39:06 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 81119-06 for ; Mon, 24 Nov 2003 09:38:38 -0400 (AST) Received: from mta11.adelphia.net (mta11.adelphia.net [68.168.78.205]) by svr1.postgresql.org (Postfix) with ESMTP id D8E29D1DA2E for ; Mon, 24 Nov 2003 09:38:28 -0400 (AST) Received: from potentialtech.com ([68.68.113.33]) by mta11.adelphia.net (InterMail vM.5.01.06.05 201-253-122-130-105-20030824) with ESMTP id <20031124133835.QRGN1464.mta11.adelphia.net@potentialtech.com>; Mon, 24 Nov 2003 08:38:35 -0500 Message-ID: <3FC209D7.5080303@potentialtech.com> Date: Mon, 24 Nov 2003 08:38:31 -0500 From: Bill Moran User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.4) Gecko/20031005 X-Accept-Language: en-us, en MIME-Version: 1.0 To: William Yu Cc: pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? References: In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/207 X-Sequence-Number: 4767 William Yu wrote: > My situation is this. We have a semi-production server where we > pre-process data and then upload the finished data to our production > servers. We need the fastest possible write performance. Having the DB > go corrupt due to power loss/OS crash is acceptable because we can > always restore from last night and re-run everything that was done since > then. > > I already have fsync off. Short of buying more hardware -- which I will > probably do anyways once I figure out whether I need more CPU, memory or > disk -- what else can I do to max out the speed? Operation mix is about > 50% select, 40% insert, 10% update. In line with what Tom Lane said, you may want to look at the various memory databases available (I'm not familiar with any one to recommend, though) If you can fit the whole database in RAM, that would work great, if not, you may be able to split the DB up and put the most used tables just in the memory database. I have also seen a number tutorials on how to put a database on a RAM disk. This helps, but it's still not as fast as a database server that's designed to keep all its data in RAM. -- Bill Moran Potential Technologies http://www.potentialtech.com From pgsql-performance-owner@postgresql.org Mon Nov 24 13:14:51 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3AD5DD1D896 for ; Mon, 24 Nov 2003 17:14:47 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 70645-07 for ; Mon, 24 Nov 2003 13:14:15 -0400 (AST) Received: from trade-india.com (ns5.trade-india.com [66.234.10.13]) by svr1.postgresql.org (Postfix) with SMTP id 50841D1D86A for ; Mon, 24 Nov 2003 13:14:14 -0400 (AST) Received: (qmail 18307 invoked from network); 24 Nov 2003 17:14:58 -0000 Received: from unknown (HELO busybox.trade-india.com) (203.145.130.142) by ns5.trade-india.com with SMTP; 24 Nov 2003 17:14:58 -0000 From: Rajesh Kumar Mallah Organization: Trade-India.com To: pgsql-performance@postgresql.org Subject: VACUUM problems with 7.4 Date: Mon, 24 Nov 2003 22:43:59 +0530 User-Agent: KMail/1.5.3 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311242243.59296.mallah@trade-india.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/208 X-Sequence-Number: 4768 I am sure there is no transaction open with the table banner_stats2. Still VACUUM FULL does not seems to effective in removing the dead rows. Can any one please help? Regds mallah tradein_clients=# VACUUM FULL verbose banner_stats2 ; INFO: vacuuming "public.banner_stats2" INFO: "banner_stats2": found 0 removable, 741912 nonremovable row versions in 6710 pages DETAIL: 737900 dead row versions cannot be removed yet. Nonremovable row versions range from 61 to 72 bytes long. There were 120 unused item pointers. Total free space (including removable row versions) is 246672 bytes. 0 pages are or will become empty, including 0 at the end of the table. 557 pages containing 61344 free bytes are potential move destinations. CPU 0.15s/1.23u sec elapsed 1.38 sec. INFO: index "banner_stats_pkey" now contains 741912 row versions in 2173 pages DETAIL: 0 index row versions were removed. 0 index pages have been deleted, 0 are currently reusable. CPU 0.03s/0.05u sec elapsed 0.09 sec. INFO: "banner_stats2": moved 0 row versions, truncated 6710 to 6710 pages DETAIL: CPU 0.00s/0.00u sec elapsed 0.00 sec. VACUUM tradein_clients=# From pgsql-performance-owner@postgresql.org Mon Nov 24 13:21:45 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1674DD1B50A for ; Mon, 24 Nov 2003 17:21:42 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 74973-10 for ; Mon, 24 Nov 2003 13:21:10 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 678ECD1C914 for ; Mon, 24 Nov 2003 13:21:10 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hAOHLANu001635 for ; Mon, 24 Nov 2003 17:21:10 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hAOHJLOD001134 for pgsql-performance@postgresql.org; Mon, 24 Nov 2003 17:19:21 GMT From: William Yu X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Maximum Possible Insert Performance? Date: Mon, 24 Nov 2003 09:19:20 -0800 Organization: Hub.Org Networking Services Lines: 16 Message-ID: References: <3FC19C2C.4050704@myrealbox.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Complaints-To: usenet@news.hub.org User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en-us, en In-Reply-To: <3FC19C2C.4050704@myrealbox.com> To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/209 X-Sequence-Number: 4769 This is an intriguing thought which leads me to think about a similar solution for even a production server and that's a solid state drive for just the WAL. What's the max disk space the WAL would ever take up? There's quite a few 512MB/1GB/2GB solid state drives available now in the ~$200-$500 range and if you never hit those limits... When my current job batch is done, I'll save a copy of the dir and give the WAL on ramdrive a test. And perhaps even buy a Sandisk at the local store and run that through the hooper. Shridhar Daithankar wrote: > > Mount WAL on RAM disk. WAL is most often hit area for heavy > updates/inserts. If you spped that up, things should be pretty faster. From pgsql-performance-owner@postgresql.org Mon Nov 24 13:27:30 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9DC62D1C943 for ; Mon, 24 Nov 2003 17:27:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 76321-06 for ; Mon, 24 Nov 2003 13:26:56 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 4A961D1B531 for ; Mon, 24 Nov 2003 13:26:55 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3969930; Mon, 24 Nov 2003 09:27:42 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: William Yu , pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? Date: Mon, 24 Nov 2003 09:25:58 -0800 User-Agent: KMail/1.4.3 References: <3FC19C2C.4050704@myrealbox.com> In-Reply-To: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311240925.58070.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/210 X-Sequence-Number: 4770 William, > When my current job batch is done, I'll save a copy of the dir and give > the WAL on ramdrive a test. And perhaps even buy a Sandisk at the local > store and run that through the hooper. We'll be interested in the results. The Sandisk won't be much of a performance test; last I checked, their access speed was about 1/2 that of a fast SCSI drive. But it could be a feasability test for the more expensive RAMdrive approach. -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Mon Nov 24 13:45:06 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 183A9D1CCE0 for ; Mon, 24 Nov 2003 17:45:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 75930-10 for ; Mon, 24 Nov 2003 13:44:31 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id B0AECD1C96C for ; Mon, 24 Nov 2003 13:44:30 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAOHiU19021400; Mon, 24 Nov 2003 12:44:30 -0500 (EST) To: Rajesh Kumar Mallah Cc: pgsql-performance@postgresql.org Subject: Re: VACUUM problems with 7.4 In-reply-to: <200311242243.59296.mallah@trade-india.com> References: <200311242243.59296.mallah@trade-india.com> Comments: In-reply-to Rajesh Kumar Mallah message dated "Mon, 24 Nov 2003 22:43:59 +0530" Date: Mon, 24 Nov 2003 12:44:30 -0500 Message-ID: <21399.1069695870@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/211 X-Sequence-Number: 4771 Rajesh Kumar Mallah writes: > I am sure there is no transaction open with the table banner_stats2. > Still VACUUM FULL does not seems to effective in removing the > dead rows. That is not the issue --- the limiting factor is what is your oldest open transaction, period. Whether it has yet looked at this table is not relevant, because the system has no way to know whether it might decide to do so later. regards, tom lane From pgsql-performance-owner@postgresql.org Mon Nov 24 13:51:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 70D6AD1D453 for ; Mon, 24 Nov 2003 17:51:42 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 76764-06 for ; Mon, 24 Nov 2003 13:51:11 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 73A87D1B50A for ; Mon, 24 Nov 2003 13:51:10 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hAOHp9Nu007323 for ; Mon, 24 Nov 2003 17:51:09 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hAOHjOdR006303 for pgsql-performance@postgresql.org; Mon, 24 Nov 2003 17:45:24 GMT From: William Yu X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Maximum Possible Insert Performance? Date: Mon, 24 Nov 2003 09:45:22 -0800 Organization: Hub.Org Networking Services Lines: 22 Message-ID: References: <3FC19C2C.4050704@myrealbox.com> <200311240925.58070.josh@agliodbs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Complaints-To: usenet@news.hub.org User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en-us, en In-Reply-To: <200311240925.58070.josh@agliodbs.com> To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/212 X-Sequence-Number: 4772 Josh Berkus wrote: > William, > > >>When my current job batch is done, I'll save a copy of the dir and give >>the WAL on ramdrive a test. And perhaps even buy a Sandisk at the local >>store and run that through the hooper. > > > We'll be interested in the results. The Sandisk won't be much of a > performance test; last I checked, their access speed was about 1/2 that of a > fast SCSI drive. But it could be a feasability test for the more expensive > RAMdrive approach. > The SanDisks do seem a bit pokey at 16MBps. On the otherhand, you could get 4 of these suckers, put them in a mega-RAID-0 stripe for 64MBps. You shouldn't need to do mirroring with a solid state drive. Time to Google up some more solid state drive vendors. From pgsql-performance-owner@postgresql.org Mon Nov 24 14:02:30 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 74F0FD1C914 for ; Mon, 24 Nov 2003 18:02:29 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 80820-07 for ; Mon, 24 Nov 2003 14:01:58 -0400 (AST) Received: from trade-india.com (ns1.trade-india.com [66.234.10.14]) by svr1.postgresql.org (Postfix) with SMTP id 85BF1D1C976 for ; Mon, 24 Nov 2003 14:01:57 -0400 (AST) Received: (qmail 10900 invoked from network); 24 Nov 2003 18:04:06 -0000 Received: from ns1.trade-india.com (HELO trade-india.com) (66.234.10.14) by ns1.trade-india.com with SMTP; 24 Nov 2003 18:04:06 -0000 Received: from 203.145.130.142 (SquirrelMail authenticated user mallah) by mail.trade-india.com with HTTP; Mon, 24 Nov 2003 23:34:06 +0530 (IST) Message-ID: <33398.203.145.130.142.1069697046.squirrel@mail.trade-india.com> Date: Mon, 24 Nov 2003 23:34:06 +0530 (IST) Subject: Re: VACUUM problems with 7.4 From: To: In-Reply-To: <21399.1069695870@sss.pgh.pa.us> References: <200311242243.59296.mallah@trade-india.com> <21399.1069695870@sss.pgh.pa.us> X-Priority: 3 Importance: Normal X-MSMail-Priority: Normal Cc: X-Mailer: SquirrelMail (version 1.2.6) MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/213 X-Sequence-Number: 4773 > Rajesh Kumar Mallah writes: >> I am sure there is no transaction open with the table banner_stats2. Still VACUUM FULL does >> not seems to effective in removing the >> dead rows. > > That is not the issue --- the limiting factor is what is your oldest open transaction, period. > Whether it has yet looked at this table is not relevant, because the system has no way to know > whether it might decide to do so later. Ok , shutting down the database and vacumming immediatly after starting helped . But it was not this bad in 7.3 as far as i understand. Is it something that has come up in 7.4 only , if so any solution to this issue? BTW can you please tell me if its safe to upgrade from RC2 to 7.4 final without initdb? [ i am still on RC2 :( ] Regds Mallah. AFTER RESTARTING DATABASE: tradein_clients=# VACUUM FULL verbose banner_stats2 ; INFO: vacuuming "public.banner_stats2" INFO: "banner_stats2": found 737900 removable, 4012 nonremovable row versions in 6710 pages DETAIL: 0 dead row versions cannot be removed yet. Nonremovable row versions range from 61 to 72 bytes long. There were 120 unused item pointers. Total free space (including removable row versions) is 51579272 bytes. 6387 pages are or will become empty, including 0 at the end of the table. 6686 pages containing 51578312 free bytes are potential move destinations. CPU 0.17s/0.09u sec elapsed 0.26 sec. INFO: index "banner_stats_pkey" now contains 4012 row versions in 2165 pages DETAIL: 737900 index row versions were removed. 1813 index pages have been deleted, 1813 are currently reusable. CPU 0.16s/1.58u sec elapsed 1.97 sec. INFO: "banner_stats2": moved 785 row versions, truncated 6710 to 38 pages DETAIL: CPU 0.17s/0.54u sec elapsed 8.30 sec. INFO: index "banner_stats_pkey" now contains 4012 row versions in 2165 pages DETAIL: 785 index row versions were removed. 1821 index pages have been deleted, 1821 are currently reusable. CPU 0.00s/0.02u sec elapsed 0.50 sec. VACUUM tradein_clients=# tradein_clients=# VACUUM FULL verbose banner_stats2 ; INFO: vacuuming "public.banner_stats2" INFO: "banner_stats2": found 0 removable, 4012 nonremovable row versions in 38 pages DETAIL: 0 dead row versions cannot be removed yet. Nonremovable row versions range from 61 to 72 bytes long. There were 100 unused item pointers. Total free space (including removable row versions) is 7368 bytes. 0 pages are or will become empty, including 0 at the end of the table. 2 pages containing 5984 free bytes are potential move destinations. CPU 0.00s/0.00u sec elapsed 0.00 sec. INFO: index "banner_stats_pkey" now contains 4012 row versions in 2165 pages DETAIL: 0 index row versions were removed. 1821 index pages have been deleted, 1821 are currently reusable. CPU 0.00s/0.00u sec elapsed 0.00 sec. INFO: "banner_stats2": moved 0 row versions, truncated 38 to 38 pages DETAIL: CPU 0.00s/0.00u sec elapsed 0.00 sec. VACUUM tradein_clients=# > > regards, tom lane > > ---------------------------(end of broadcast)--------------------------- TIP 2: you can get off > all lists at once with the unregister command > (send "unregister YourEmailAddressHere" to majordomo@postgresql.org) ----------------------------------------- Over 1,00,000 exporters are waiting for your order! Click below to get in touch with leading Indian exporters listed in the premier trade directory Exporters Yellow Pages. http://www.trade-india.com/dyn/gdh/eyp/ From pgsql-performance-owner@postgresql.org Mon Nov 24 14:06:08 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id BEC6AD1CB08 for ; Mon, 24 Nov 2003 18:06:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78108-09 for ; Mon, 24 Nov 2003 14:05:36 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id C3689D1C96C for ; Mon, 24 Nov 2003 14:05:35 -0400 (AST) Received: from [63.195.55.98] (HELO spooky) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3970123; Mon, 24 Nov 2003 10:06:23 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Organization: Aglio Database Solutions To: William Yu , pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? Date: Mon, 24 Nov 2003 10:04:37 -0800 User-Agent: KMail/1.4.3 References: <200311240925.58070.josh@agliodbs.com> In-Reply-To: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Message-Id: <200311241004.37969.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/214 X-Sequence-Number: 4774 William, > The SanDisks do seem a bit pokey at 16MBps. On the otherhand, you could > get 4 of these suckers, put them in a mega-RAID-0 stripe for 64MBps. You > shouldn't need to do mirroring with a solid state drive. I wouldn't count on RAID0 improving the speed of SANDisk's much. How are you connecting to them? USB? USB doesn't support fast parallel data access. Now, if it turns out that 256MB ramdisks are less than 1/5 the cost of 1GB ramdisks, then that's worth considering. You're right, though, mirroring a solid state drive is pretty pointless; if power fails, both mirrors are dead. As I said before, though, we're all very interested in this test. Using a ramdisk for WAL has been discussed on this list numerous times but not attempted by anyone who published their results. All that aside, though, I think you should also experiment with the Background Writer patch recently discussed on Hackers, as it may give you a performance boost as well. -- Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Mon Nov 24 14:52:19 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id AC9A9D1D8A9 for ; Mon, 24 Nov 2003 18:52:09 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 99120-02 for ; Mon, 24 Nov 2003 14:51:39 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 66847D1D880 for ; Mon, 24 Nov 2003 14:51:10 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hAOIpANu018836 for ; Mon, 24 Nov 2003 18:51:10 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hAOINbCO013502 for pgsql-performance@postgresql.org; Mon, 24 Nov 2003 18:23:37 GMT From: William Yu X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Maximum Possible Insert Performance? Date: Mon, 24 Nov 2003 10:23:36 -0800 Organization: Hub.Org Networking Services Lines: 29 Message-ID: References: <200311240925.58070.josh@agliodbs.com> <200311241004.37969.josh@agliodbs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Complaints-To: usenet@news.hub.org User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en-us, en In-Reply-To: <200311241004.37969.josh@agliodbs.com> To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/215 X-Sequence-Number: 4775 Josh Berkus wrote: > William, > > >>The SanDisks do seem a bit pokey at 16MBps. On the otherhand, you could >>get 4 of these suckers, put them in a mega-RAID-0 stripe for 64MBps. You >>shouldn't need to do mirroring with a solid state drive. > > > I wouldn't count on RAID0 improving the speed of SANDisk's much. How are you > connecting to them? USB? USB doesn't support fast parallel data access. You can get ATA SanDisks up to 2GB. Another vendor I checked out -- BitMicro -- has solid state drives for SATA, SCSI and FiberChannel. I'd definitely would not use USB SSDs -- USB performance would be so pokey to be useless. > Now, if it turns out that 256MB ramdisks are less than 1/5 the cost of 1GB > ramdisks, then that's worth considering. Looks like they're linear with size. SanDisk Flashdrive 1GB is about $1000 while 256MB is $250. > You're right, though, mirroring a solid state drive is pretty pointless; if > power fails, both mirrors are dead. Actually no. Solid state memory is non-volatile. They retain data even without power. From pgsql-performance-owner@postgresql.org Mon Nov 24 15:30:35 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E598ED1D453 for ; Mon, 24 Nov 2003 19:30:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 14012-02 for ; Mon, 24 Nov 2003 15:30:04 -0400 (AST) Received: from anchor-post-35.mail.demon.net (anchor-post-35.mail.demon.net [194.217.242.85]) by svr1.postgresql.org (Postfix) with ESMTP id C6FE9D1D643 for ; Mon, 24 Nov 2003 15:29:45 -0400 (AST) Received: from mwynhau.demon.co.uk ([193.237.186.96] helo=mainbox.archonet.com) by anchor-post-35.mail.demon.net with esmtp (Exim 3.35 #1) id 1AOMP2-000DtI-0Z; Mon, 24 Nov 2003 19:29:44 +0000 Received: by mainbox.archonet.com (Postfix, from userid 529) id 90E4C17CEC; Mon, 24 Nov 2003 19:29:42 +0000 (GMT) Received: from client17.archonet.com (client17.archonet.com [192.168.1.17]) by mainbox.archonet.com (Postfix) with ESMTP id EBF861622E; Mon, 24 Nov 2003 19:29:40 +0000 (GMT) From: Richard Huxton To: stephen farrell , pgsql-performance@postgresql.org Subject: Re: Problem with insert into select... Date: Mon, 24 Nov 2003 19:29:40 +0000 User-Agent: KMail/1.5 References: <3FBD2C5D.8000706@almaden.ibm.com> In-Reply-To: <3FBD2C5D.8000706@almaden.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311241929.40600.dev@archonet.com> X-Bogosity: No, tests=bogofilter, spamicity=0.000000, version=0.15.3 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/223 X-Sequence-Number: 4783 On Thursday 20 November 2003 21:04, stephen farrell wrote: > I'm having a problem with a queyr like: INSERT INTO FACT (x,x,x,x,x,x) > SELECT a.key,b.key,c.key,d.key,e.key,f.key from x,a,b,c,d,e,f where x=a > and x=b .... -- postgres7.4 is running out of memory. When this has happened to me it's always been because I've got an unconstrained join due to pilot error. Try an EXPLAIN on the select part and see if that pops up anything. -- Richard Huxton Archonet Ltd From pgsql-performance-owner@postgresql.org Mon Nov 24 17:20:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5C373D1D609 for ; Mon, 24 Nov 2003 21:20:45 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 37148-07 for ; Mon, 24 Nov 2003 17:20:16 -0400 (AST) Received: from smtp.noos.fr (nan-smtp-20.noos.net [212.198.2.120]) by svr1.postgresql.org (Postfix) with ESMTP id 87615D1D8AF for ; Mon, 24 Nov 2003 17:20:13 -0400 (AST) Received: (qmail 7192 invoked by uid 0); 24 Nov 2003 21:20:14 -0000 Received: from unknown (HELO bigfoot.com) ([212.198.40.101]) (envelope-sender ) by 212.198.2.120 (qmail-ldap-1.03) with SMTP for ; 24 Nov 2003 21:20:14 -0000 Message-ID: <3FC267FB.806@bigfoot.com> Date: Mon, 24 Nov 2003 21:20:11 +0100 From: Gaetano Mendola User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "pgsql-performance@postgresql.org" Cc: Torsten Schulz Subject: Re: Optimize References: <3FC26E96.9090101@be-a-part.de> In-Reply-To: <3FC26E96.9090101@be-a-part.de> X-Enigmail-Version: 0.81.7.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/225 X-Sequence-Number: 4785 Torsten Schulz wrote: > Yes, I know: very difficult question, but I don't know what to do now. > > Our Server: > Dual-CPU with 1.2 GHz > 1.5 GB RAM > > Our Problem: We are a Community. Between 19 and 21 o clock we have >350 > User in the Community. But then, the Database are very slow. And we have > per CPU ~20-30% idle-time. May we know the postgres version that you are running and see the query that run slow ? Is also usefull take a look at your postgresql configuration. You can see doing select * from pg_stat_activity the queries that are currently running on your server, and do a explain analize on it to see which one is the bottleneck. If you are running the 7.4 you can see on the log the total ammount for each query. Let us know. Regards Gaetano Mendola From pgsql-performance-owner@postgresql.org Mon Nov 24 16:48:52 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5D70AD1D8AF for ; Mon, 24 Nov 2003 20:48:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 31109-07 for ; Mon, 24 Nov 2003 16:48:24 -0400 (AST) Received: from natsmtp01.rzone.de (natsmtp01.rzone.de [81.169.145.166]) by svr1.postgresql.org (Postfix) with ESMTP id B59A9D1B531 for ; Mon, 24 Nov 2003 16:48:21 -0400 (AST) Received: from be-a-part.de (pD9E820FC.dip.t-dialin.net [217.232.32.252]) by post.webmailer.de (8.12.10/8.12.10) with ESMTP id hAOKmMMk010850 for ; Mon, 24 Nov 2003 21:48:22 +0100 (MET) Message-ID: <3FC26E96.9090101@be-a-part.de> Date: Mon, 24 Nov 2003 21:48:22 +0100 From: Torsten Schulz User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5b) Gecko/20030827 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Optimize Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/224 X-Sequence-Number: 4784 Yes, I know: very difficult question, but I don't know what to do now. Our Server: Dual-CPU with 1.2 GHz 1.5 GB RAM Our Problem: We are a Community. Between 19 and 21 o clock we have >350 User in the Community. But then, the Database are very slow. And we have per CPU ~20-30% idle-time. Has anyone an idea what's the best configuration for thta server? Many Greetings T. Schulz (with very bad english, i know) From pgsql-performance-owner@postgresql.org Mon Nov 24 17:56:51 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2D505D1D643 for ; Mon, 24 Nov 2003 21:56:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 49023-06 for ; Mon, 24 Nov 2003 17:56:21 -0400 (AST) Received: from smtp.noos.fr (nan-smtp-16.noos.net [212.198.2.124]) by svr1.postgresql.org (Postfix) with ESMTP id AD30AD1B520 for ; Mon, 24 Nov 2003 17:56:18 -0400 (AST) Received: (qmail 13334 invoked by uid 0); 24 Nov 2003 21:56:20 -0000 Received: from unknown (HELO bigfoot.com) ([212.198.40.101]) (envelope-sender ) by 212.198.2.124 (qmail-ldap-1.03) with SMTP for ; 24 Nov 2003 21:56:20 -0000 Message-ID: <3FC27071.9000001@bigfoot.com> Date: Mon, 24 Nov 2003 21:56:17 +0100 From: Gaetano Mendola User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "pgsql-performance@postgresql.org" Cc: Torsten Schulz Subject: Re: Optimize References: <3FC26E96.9090101@be-a-part.de> <3FC267FB.806@bigfoot.com> <3FC27911.7070403@be-a-part.de> In-Reply-To: <3FC27911.7070403@be-a-part.de> X-Enigmail-Version: 0.81.7.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/228 X-Sequence-Number: 4788 Torsten Schulz wrote: > Gaetano Mendola wrote: > >> Torsten Schulz wrote: >> >>> Yes, I know: very difficult question, but I don't know what to do now. >>> >>> Our Server: >>> Dual-CPU with 1.2 GHz >>> 1.5 GB RAM >>> >>> Our Problem: We are a Community. Between 19 and 21 o clock we have >>> >350 User in the Community. But then, the Database are very slow. And >>> we have per CPU ~20-30% idle-time. >> >> >> >> May we know the postgres version that you are running and >> see the query that run slow ? > > > Postgres: 7.3.2 > Query: All queries > > Configuration: > max_connections = 1000 # Must be, if lower then 500 we become > connection-errors > shared_buffers = 5000 # 2*max_connections, min 16 > max_fsm_relations = 1000 # min 10, fsm is free space map > max_fsm_pages = 2000000 # min 1000, fsm is free space map > max_locks_per_transaction = 64 # min 10 > wal_buffers = 2000 # min 4 > > sort_mem = 32768 # min 32 > vacuum_mem = 32768 # min 1024 > > fsync = false > > enable_seqscan = true > enable_indexscan = true > enable_tidscan = true > enable_sort = true > enable_nestloop = true > enable_mergejoin = true > enable_hashjoin = true > > effective_cache_size = 96000 # default in 8k pages With 500 connection at the sime time 32MB for sort_mem can be too much. What say "iostat 1" and "vmstat 1" ? Try also to reduce this costs: random_page_cost = 2.5 cpu_tuple_cost = 0.005 cpu_index_tuple_cost = 0.0005 BTW take a query and show us the result of explain analyze. Regards Gaetano Mendola From pgsql-performance-owner@postgresql.org Mon Nov 24 17:33:38 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id BF53BD1D9D3 for ; Mon, 24 Nov 2003 21:33:36 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 43164-07 for ; Mon, 24 Nov 2003 17:33:07 -0400 (AST) Received: from natsmtp01.rzone.de (natsmtp01.rzone.de [81.169.145.166]) by svr1.postgresql.org (Postfix) with ESMTP id 0FCDAD1D8EA for ; Mon, 24 Nov 2003 17:33:05 -0400 (AST) Received: from be-a-part.de (pD9E820FC.dip.t-dialin.net [217.232.32.252]) by post.webmailer.de (8.12.10/8.12.10) with ESMTP id hAOLX5Al021454 for ; Mon, 24 Nov 2003 22:33:06 +0100 (MET) Message-ID: <3FC27911.7070403@be-a-part.de> Date: Mon, 24 Nov 2003 22:33:05 +0100 From: Torsten Schulz User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5b) Gecko/20030827 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Re: Optimize References: <3FC26E96.9090101@be-a-part.de> <3FC267FB.806@bigfoot.com> In-Reply-To: <3FC267FB.806@bigfoot.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/226 X-Sequence-Number: 4786 Gaetano Mendola wrote: > Torsten Schulz wrote: > >> Yes, I know: very difficult question, but I don't know what to do now. >> >> Our Server: >> Dual-CPU with 1.2 GHz >> 1.5 GB RAM >> >> Our Problem: We are a Community. Between 19 and 21 o clock we have >> >350 User in the Community. But then, the Database are very slow. And >> we have per CPU ~20-30% idle-time. > > > May we know the postgres version that you are running and > see the query that run slow ? Postgres: 7.3.2 Query: All queries Configuration: max_connections = 1000 # Must be, if lower then 500 we become connection-errors shared_buffers = 5000 # 2*max_connections, min 16 max_fsm_relations = 1000 # min 10, fsm is free space map max_fsm_pages = 2000000 # min 1000, fsm is free space map max_locks_per_transaction = 64 # min 10 wal_buffers = 2000 # min 4 sort_mem = 32768 # min 32 vacuum_mem = 32768 # min 1024 fsync = false enable_seqscan = true enable_indexscan = true enable_tidscan = true enable_sort = true enable_nestloop = true enable_mergejoin = true enable_hashjoin = true effective_cache_size = 96000 # default in 8k pages That are all uncommented lines. I've found the values in internet and had tested it. But in performance are no difference between old configuration an this. > Is also usefull take a look at your postgresql configuration. > You can see doing select * from pg_stat_activity the > queries that are currently running on your server, and > do a explain analize on it to see which one is the > bottleneck. If you are running the 7.4 you can see on > the log the total ammount for each query. > I'll show tomorrow for this, today it is too late, the performance is now perfect. It's only slow on this 2 hours with so many users on server. Oh, and i can't update to 7.4. The Chat don't run with libraries of 7.4 From pgsql-performance-owner@postgresql.org Mon Nov 24 17:46:21 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 09166D1D86B for ; Mon, 24 Nov 2003 21:46:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 48630-02 for ; Mon, 24 Nov 2003 17:45:51 -0400 (AST) Received: from zigo.dhs.org (as2-4-3.an.g.bonet.se [194.236.34.191]) by svr1.postgresql.org (Postfix) with ESMTP id 019D6D1D643 for ; Mon, 24 Nov 2003 17:45:48 -0400 (AST) Received: from zigo.dhs.org (zigo [127.0.0.1]) by zigo.dhs.org (8.12.8/8.12.8) with ESMTP id hAOLjiL2028737; Mon, 24 Nov 2003 22:45:44 +0100 Received: from localhost (db@localhost) by zigo.dhs.org (8.12.8/8.12.8/Submit) with ESMTP id hAOLjioS028733; Mon, 24 Nov 2003 22:45:44 +0100 Date: Mon, 24 Nov 2003 22:45:44 +0100 (CET) From: Dennis Bjorklund To: Torsten Schulz Cc: pgsql-performance@postgresql.org Subject: Re: Optimize In-Reply-To: <3FC27911.7070403@be-a-part.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/227 X-Sequence-Number: 4787 On Mon, 24 Nov 2003, Torsten Schulz wrote: > sort_mem = 32768 # min 32 32 meg per sort can be a lot in total if you have many clients sorting things. I assume you have checked so that the computer is not pushed into swapping when you have the peak with lots of users. A swapping computer is never fast. Using some swap space is not bad, but a lot of page in and page out to the swap is not good. -- /Dennis From pgsql-performance-owner@postgresql.org Mon Nov 24 18:03:20 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0E4B2D1C92F for ; Mon, 24 Nov 2003 22:03:20 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 52085-01 for ; Mon, 24 Nov 2003 18:02:51 -0400 (AST) Received: from sccrmhc13.comcast.net (sccrmhc13.comcast.net [204.127.202.64]) by svr1.postgresql.org (Postfix) with ESMTP id B66CBD1D643 for ; Mon, 24 Nov 2003 18:02:48 -0400 (AST) Received: from game (12-250-179-10.client.attbi.com[12.250.179.10]) by comcast.net (sccrmhc13) with SMTP id <2003112422025001600etecle>; Mon, 24 Nov 2003 22:02:50 +0000 Message-ID: <2b1f01c3b2d6$c403ba70$0201a8c0@game> Reply-To: "MK Spam" From: "MK Spam" To: Subject: Where to start for performance problem? Date: Mon, 24 Nov 2003 16:03:17 -0600 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_2B1C_01C3B2A4.7921BA20" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=1.2 tagged_above=0.0 required=5.0 tests=HTML_10_20 X-Spam-Level: * X-Archive-Number: 200311/229 X-Sequence-Number: 4789 This is a multi-part message in MIME format. ------=_NextPart_000_2B1C_01C3B2A4.7921BA20 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable I've scanned some of the archives and have learned a lot about different pe= rformance tuning practices. I will be looking into using many of these ide= as but I'm not sure they address the issue I am currently experiencing. First, I'm a total newb with postgresql. Like many before me, I have inher= ited many responsibilities outside of my original job description due to la= yoffs. I am now the sole developer/support for a software product. *sigh*= This product uses postgresql. I am familiar with the basics of sql and h= ave worked on the database and code for the software but by no means am I p= roficient with postgresql. The archives of this list provides many ideas for improving performance, bu= t the problem we are having is gradually degrading performance ending in po= stgres shutting down. So it's not a matter of optimizing a complex query t= o take 5 seconds instead of 60 seconds. From what I can tell we are using = the VACUUM command on a schedule but it doesn't seem to prevent the databas= e from becoming "congested" as we refer to it. :] Anyway, the only way I = know to "fix" the problem is to export (pg_dump) the db, drop the database,= recreate the database and import the dump. This seems to return performan= ce back to normal but obviously isn't a very good "solution". The slowdown= and subsequent crash can take as little as 1 week for databases with a lot= of data or go as long as a few weeks to a month for smaller data sets. I don't really know where to start looking for a solution. Any advice on w= here to start, understanding that I am a newb, would be greatly appreciated= . Thank you. Nid ------=_NextPart_000_2B1C_01C3B2A4.7921BA20 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
I've scanned some of the archives and have= learned=20 a lot about different performance tuning practices.  I will be looking= into=20 using many of these ideas but I'm not sure they address the issue I am curr= ently=20 experiencing.
 
First, I'm a total newb with postgresql.&n= bsp; Like=20 many before me, I have inherited many responsibilities outside of my origin= al=20 job description due to layoffs.  I am now the sole developer/support f= or a=20 software product.  *sigh*  This product uses postgresql.  I = am=20 familiar with the basics of sql and have worked on the database and code fo= r the=20 software but by no means am I proficient with postgresql.
 
The archives of this list provides many id= eas for=20 improving performance, but the problem we are having is gradually degrading= =20 performance ending in postgres shutting down.  So it's not a matter of= =20 optimizing a complex query to take 5 seconds instead of 60 seconds.&nb= sp;=20 >From what I can tell we are using the VACUUM command on a schedule but it= =20 doesn't seem to prevent the database from becoming "congested" as we refer = to=20 it.  :]  Anyway, the only way I know to "fix" the problem is to e= xport=20 (pg_dump) the db, drop the database, recreate the database and import the= =20 dump.  This seems to return performance back to normal but obviously i= sn't=20 a very good "solution".  The slowdown and subsequent crash can take as= =20 little as 1 week for databases with a lot of data or go as long as a few we= eks=20 to a month for smaller data sets.
 
I don't really know where to start looking= for a=20 solution.  Any advice on where to start, understanding that I am a new= b,=20 would be greatly appreciated.  Thank you.
 
Nid
------=_NextPart_000_2B1C_01C3B2A4.7921BA20-- From pgsql-performance-owner@postgresql.org Mon Nov 24 18:08:04 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 38570D1D80D for ; Mon, 24 Nov 2003 22:08:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 50487-04 for ; Mon, 24 Nov 2003 18:07:34 -0400 (AST) Received: from sccrmhc11.comcast.net (sccrmhc11.comcast.net [204.127.202.55]) by svr1.postgresql.org (Postfix) with ESMTP id 00DD8D1B520 for ; Mon, 24 Nov 2003 18:07:32 -0400 (AST) Received: from game (12-250-179-10.client.attbi.com[12.250.179.10]) by comcast.net (sccrmhc11) with SMTP id <2003112422073301100nq75ae>; Mon, 24 Nov 2003 22:07:33 +0000 Message-ID: <2b4b01c3b2d7$6ce8e160$0201a8c0@game> Reply-To: "Nid" From: "Nid" To: References: <2b1f01c3b2d6$c403ba70$0201a8c0@game> Subject: Re: Where to start for performance problem? Date: Mon, 24 Nov 2003 16:08:00 -0600 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_2B48_01C3B2A5.22211FD0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=1.5 tagged_above=0.0 required=5.0 tests=HTML_20_30, REFERENCES X-Spam-Level: * X-Archive-Number: 200311/230 X-Sequence-Number: 4790 This is a multi-part message in MIME format. ------=_NextPart_000_2B48_01C3B2A5.22211FD0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable My apologies for the "From" name of MK Spam. That references an email acco= unt I made for signing up for things on the net. :] Nid ----- Original Message -----=20 From: MK Spam=20 To: pgsql-performance@postgresql.org=20 Sent: Monday, November 24, 2003 4:03 PM Subject: [PERFORM] Where to start for performance problem? I've scanned some of the archives and have learned a lot about different = performance tuning practices. I will be looking into using many of these i= deas but I'm not sure they address the issue I am currently experiencing. First, I'm a total newb with postgresql. Like many before me, I have inh= erited many responsibilities outside of my original job description due to = layoffs. I am now the sole developer/support for a software product. *sig= h* This product uses postgresql. I am familiar with the basics of sql and= have worked on the database and code for the software but by no means am I= proficient with postgresql. The archives of this list provides many ideas for improving performance, = but the problem we are having is gradually degrading performance ending in = postgres shutting down. So it's not a matter of optimizing a complex query= to take 5 seconds instead of 60 seconds. From what I can tell we are usin= g the VACUUM command on a schedule but it doesn't seem to prevent the datab= ase from becoming "congested" as we refer to it. :] Anyway, the only way = I know to "fix" the problem is to export (pg_dump) the db, drop the databas= e, recreate the database and import the dump. This seems to return perform= ance back to normal but obviously isn't a very good "solution". The slowdo= wn and subsequent crash can take as little as 1 week for databases with a l= ot of data or go as long as a few weeks to a month for smaller data sets. I don't really know where to start looking for a solution. Any advice on= where to start, understanding that I am a newb, would be greatly appreciat= ed. Thank you. Nid ------=_NextPart_000_2B48_01C3B2A5.22211FD0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
My apologies for the "From" name of MK Spa= m. =20 That references an email account I made for signing up for things on the=20 net.  :]
 
Nid
----- Original Message -----
Fro= m:=20 MK Spa= m=20
To: pgsql-performance@postgr= esql.org=20
Sent: Monday, November 24, 2003 4:= 03=20 PM
Subject: [PERFORM] Where to start = for=20 performance problem?

I've scanned some of the archives and ha= ve=20 learned a lot about different performance tuning practices.  I will = be=20 looking into using many of these ideas but I'm not sure they address the = issue=20 I am currently experiencing.
 
First, I'm a total newb with postgresql.=  =20 Like many before me, I have inherited many responsibilities outside of my= =20 original job description due to layoffs.  I am now the sole=20 developer/support for a software product.  *sigh*  This product= uses=20 postgresql.  I am familiar with the basics of sql and have worked on= the=20 database and code for the software but by no means am I proficient with= =20 postgresql.
 
The archives of this list provides many = ideas for=20 improving performance, but the problem we are having is gradually degradi= ng=20 performance ending in postgres shutting down.  So it's not a matter = of=20 optimizing a complex query to take 5 seconds instead of 60 seconds.&= nbsp;=20 From what I can tell we are using the VACUUM command on a schedule but it= =20 doesn't seem to prevent the database from becoming "congested" as we refe= r to=20 it.  :]  Anyway, the only way I know to "fix" the problem is to= =20 export (pg_dump) the db, drop the database, recreate the database and imp= ort=20 the dump.  This seems to return performance back to normal but obvio= usly=20 isn't a very good "solution".  The slowdown and subsequent crash can= take=20 as little as 1 week for databases with a lot of data or go as long as a f= ew=20 weeks to a month for smaller data sets.
 
I don't really know where to start looki= ng for a=20 solution.  Any advice on where to start, understanding that I am a n= ewb,=20 would be greatly appreciated.  Thank you.
 
Nid
------=_NextPart_000_2B48_01C3B2A5.22211FD0-- From pgsql-performance-owner@postgresql.org Mon Nov 24 21:12:12 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D16CBD1C914 for ; Tue, 25 Nov 2003 01:12:09 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 88612-04 for ; Mon, 24 Nov 2003 21:11:41 -0400 (AST) Received: from houston.familyhealth.com.au (fhnet.arach.net.au [203.22.197.21]) by svr1.postgresql.org (Postfix) with ESMTP id A0F7DD1C951 for ; Mon, 24 Nov 2003 21:11:36 -0400 (AST) Received: from familyhealth.com.au (work-40.internal [192.168.0.40]) by houston.familyhealth.com.au (8.12.9p1/8.12.9) with ESMTP id hAP1BGoD032541; Tue, 25 Nov 2003 09:11:18 +0800 (WST) (envelope-from chriskl@familyhealth.com.au) Message-ID: <3FC2ACCF.2030303@familyhealth.com.au> Date: Tue, 25 Nov 2003 09:13:51 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: MK Spam Cc: pgsql-performance@postgresql.org Subject: Re: Where to start for performance problem? References: <2b1f01c3b2d6$c403ba70$0201a8c0@game> In-Reply-To: <2b1f01c3b2d6$c403ba70$0201a8c0@game> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/231 X-Sequence-Number: 4791 > The archives of this list provides many ideas for improving performance, > but the problem we are having is gradually degrading performance ending > in postgres shutting down. So it's not a matter of optimizing a complex > query to take 5 seconds instead of 60 seconds. >From what I can tell we > are using the VACUUM command on a schedule but it doesn't seem to > prevent the database from becoming "congested" as we refer to it. :] Our busy website has a cronjob that runs VACUUM ANALYZE once an hour (vacuumdb -a -q -z). Have you tried going 'VACUUM FULL ANALYZE' (vacuumdb -a -q -z -f) instead of a dump and reload? Chris From pgsql-performance-owner@postgresql.org Mon Nov 24 21:17:06 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 48BD7D1D8FC for ; Tue, 25 Nov 2003 01:17:05 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 88267-10 for ; Mon, 24 Nov 2003 21:16:39 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 31F89D1CAD1 for ; Mon, 24 Nov 2003 21:16:36 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id 478D93695B; Mon, 24 Nov 2003 20:16:28 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1AORoZ-0001Dr-00; Mon, 24 Nov 2003 20:16:27 -0500 To: William Yu Cc: pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? References: <200311240925.58070.josh@agliodbs.com> <200311241004.37969.josh@agliodbs.com> In-Reply-To: From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 24 Nov 2003 20:16:27 -0500 Message-ID: <87brr11c44.fsf@stark.dyndns.tv> Lines: 17 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/232 X-Sequence-Number: 4792 William Yu writes: > > You're right, though, mirroring a solid state drive is pretty pointless; if > > power fails, both mirrors are dead. > > Actually no. Solid state memory is non-volatile. They retain data even without > power. Note that flash ram only has a finite number of write cycles before it fails. On the other hand that might not be so bad for WAL which writes sequentially, you can easily calculate how close you are to the maximum. For things like heap storage or swap it's awful as you can get hot spots that get written to thousands of times before the rest of the space is used. -- greg From pgsql-performance-owner@postgresql.org Mon Nov 24 22:25:22 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 26B67D1D66C for ; Tue, 25 Nov 2003 02:25:18 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 11223-08 for ; Mon, 24 Nov 2003 22:24:50 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id 29F0BD1D880 for ; Mon, 24 Nov 2003 22:24:44 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 3F394203A; Mon, 24 Nov 2003 21:24:47 -0500 (EST) To: Torsten Schulz Cc: pgsql-performance@postgresql.org Subject: Re: Optimize From: Neil Conway In-Reply-To: <3FC26E96.9090101@be-a-part.de> (Torsten Schulz's message of "Mon, 24 Nov 2003 21:48:22 +0100") References: <3FC26E96.9090101@be-a-part.de> Date: Mon, 24 Nov 2003 21:24:45 -0500 Message-ID: <87fzgdnq1e.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/233 X-Sequence-Number: 4793 Torsten Schulz writes: > Our Server: > Dual-CPU with 1.2 GHz > 1.5 GB RAM What kind of I/O subsystem is in this machine? This is an x86 machine, right? > Has anyone an idea what's the best configuration for thta server? It is difficult to say until you provide some information on the system's state during periods of heavy traffic. BTW, in addition to the machine's hardware configuration, have you looked at tuning the queries running on PostgreSQL? What about the OS kernel? -Neil From pgsql-performance-owner@postgresql.org Mon Nov 24 22:28:13 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9724BD1DBA8 for ; Tue, 25 Nov 2003 02:28:10 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 12604-05 for ; Mon, 24 Nov 2003 22:27:43 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id 1F560D1DA94 for ; Mon, 24 Nov 2003 22:27:39 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id BAFD81D9B; Mon, 24 Nov 2003 21:27:41 -0500 (EST) To: Cc: , Subject: Re: VACUUM problems with 7.4 From: Neil Conway In-Reply-To: <33398.203.145.130.142.1069697046.squirrel@mail.trade-india.com> (mallah@trade-india.com's message of "Mon, 24 Nov 2003 23:34:06 +0530 (IST)") References: <200311242243.59296.mallah@trade-india.com> <21399.1069695870@sss.pgh.pa.us> <33398.203.145.130.142.1069697046.squirrel@mail.trade-india.com> Date: Mon, 24 Nov 2003 21:27:40 -0500 Message-ID: <87brr1npwj.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests=IN_REP_TO, RCVD_IN_OSIRUSOFT_COM, REFERENCES, USER_AGENT_GNUS_UA X-Spam-Level: X-Archive-Number: 200311/234 X-Sequence-Number: 4794 writes: > But it was not this bad in 7.3 as far as i understand. No, I believe this behavior is present in any recent release of PostgreSQL. -Neil From pgsql-performance-owner@postgresql.org Mon Nov 24 23:05:57 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id BB1A0D1C96C for ; Tue, 25 Nov 2003 03:05:56 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 16802-01 for ; Mon, 24 Nov 2003 23:05:29 -0400 (AST) Received: from sccrmhc12.comcast.net (sccrmhc12.comcast.net [204.127.202.56]) by svr1.postgresql.org (Postfix) with ESMTP id B66F6D1C976 for ; Mon, 24 Nov 2003 23:05:24 -0400 (AST) Received: from game (12-250-179-10.client.attbi.com[12.250.179.10]) by comcast.net (sccrmhc12) with SMTP id <2003112503052801200l34a2e>; Tue, 25 Nov 2003 03:05:28 +0000 Message-ID: <2b8e01c3b301$0bc2f090$0201a8c0@game> Reply-To: "Nid" From: "Nid" To: "Christopher Kings-Lynne" Cc: References: <2b1f01c3b2d6$c403ba70$0201a8c0@game> <3FC2ACCF.2030303@familyhealth.com.au> Subject: Re: Where to start for performance problem? Date: Mon, 24 Nov 2003 21:05:56 -0600 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/235 X-Sequence-Number: 4795 I've been digging around in the code and found where we are executing the VACUUM command. VACUUM ANALYZE is executed every 15 minutes. We haven't tried VACUUM FULL ANALYZE. I think I read that using FULL is a good idea once a day or something. Just doing a VACUUM ANALYZE doesn't seem to be preventing our problem. Thank you for the responses. nid ----- Original Message ----- From: "Christopher Kings-Lynne" To: "MK Spam" Cc: Sent: Monday, November 24, 2003 7:13 PM Subject: Re: [PERFORM] Where to start for performance problem? > > The archives of this list provides many ideas for improving performance, > > but the problem we are having is gradually degrading performance ending > > in postgres shutting down. So it's not a matter of optimizing a complex > > query to take 5 seconds instead of 60 seconds. >From what I can tell we > > are using the VACUUM command on a schedule but it doesn't seem to > > prevent the database from becoming "congested" as we refer to it. :] > > Our busy website has a cronjob that runs VACUUM ANALYZE once an hour > (vacuumdb -a -q -z). > > Have you tried going 'VACUUM FULL ANALYZE' (vacuumdb -a -q -z -f) > instead of a dump and reload? > > Chris > > > From pgsql-performance-owner@postgresql.org Mon Nov 24 23:16:42 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CB6FED1DBAC for ; Tue, 25 Nov 2003 03:16:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 16507-03 for ; Mon, 24 Nov 2003 23:16:13 -0400 (AST) Received: from houston.familyhealth.com.au (fhnet.arach.net.au [203.22.197.21]) by svr1.postgresql.org (Postfix) with ESMTP id CDB3AD1D888 for ; Mon, 24 Nov 2003 23:16:07 -0400 (AST) Received: from familyhealth.com.au (work-40.internal [192.168.0.40]) by houston.familyhealth.com.au (8.12.9p1/8.12.9) with ESMTP id hAP3G8oD033859; Tue, 25 Nov 2003 11:16:08 +0800 (WST) (envelope-from chriskl@familyhealth.com.au) Message-ID: <3FC2CA18.3080001@familyhealth.com.au> Date: Tue, 25 Nov 2003 11:18:48 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Nid Cc: pgsql-performance@postgresql.org Subject: Re: Where to start for performance problem? References: <2b1f01c3b2d6$c403ba70$0201a8c0@game> <3FC2ACCF.2030303@familyhealth.com.au> <2b8e01c3b301$0bc2f090$0201a8c0@game> In-Reply-To: <2b8e01c3b301$0bc2f090$0201a8c0@game> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/236 X-Sequence-Number: 4796 Nid wrote: > I've been digging around in the code and found where we are executing the > VACUUM command. VACUUM ANALYZE is executed every 15 minutes. We haven't > tried VACUUM FULL ANALYZE. I think I read that using FULL is a good idea > once a day or something. Just doing a VACUUM ANALYZE doesn't seem to be > preventing our problem. Thank you for the responses. Try upgrading to PostgreSQL 7.4 and use the new pg_autovacuum daemon. This daemon will monitor your tables and vacuum and analyze whenever necessary. Chris From pgsql-performance-owner@postgresql.org Tue Nov 25 00:21:52 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9C0D8D1D61F for ; Tue, 25 Nov 2003 04:21:45 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 27484-08 for ; Tue, 25 Nov 2003 00:21:17 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 585DAD1CAD1 for ; Tue, 25 Nov 2003 00:21:16 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hAP4LGNu050445 for ; Tue, 25 Nov 2003 04:21:16 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hAP4CTZT049017 for pgsql-performance@postgresql.org; Tue, 25 Nov 2003 04:12:29 GMT From: Christopher Browne X-Newsgroups: comp.databases.postgresql.performance Subject: Re: [HACKERS] More detail on settings for pgavd? Date: Mon, 24 Nov 2003 23:08:08 -0500 Organization: cbbrowne Computing Inc Lines: 38 Message-ID: References: <200311181558.45847.josh@agliodbs.com> <1069424054.29672.12462.camel@camel> <3FBE21D5.20508@zeut.net> <200311210909.00978.josh@agliodbs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Complaints-To: usenet@news.hub.org X-message-flag: Outlook is rather hackable, isn't it? X-Home-Page: http://www.cbbrowne.com/info/ X-Affero: http://svcs.affero.net/rm.php?r=cbbrowne User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) Cancel-Lock: sha1:XUF1rnSlb/8Axa0fuCtewYQLVvk= To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests=QUOTED_EMAIL_TEXT, RCVD_IN_OSIRUSOFT_COM, REFERENCES, USER_AGENT_GNUS_UA X-Spam-Level: X-Archive-Number: 200311/237 X-Sequence-Number: 4797 After a long battle with technology, josh@agliodbs.com (Josh Berkus), an earthling, wrote: >> As long as pg_autovacuum remains a contrib module, I don't think >> any changes to the system catelogs will be make. If pg_autovacuum >> is deemed ready to move out of contrib, then we can talk about the >> above. > > But we could create a config file that would store stuff in a > flatfile table, OR we could add our own "system table" that would be > created when one "initializes" pg_avd. The problem with introducing a "config file" is that you then have to introduce a language and a parser for that language. That introduces rather a lot of complexity. That was the BIG problem with pgavd (which is a discarded project; pg_autovacuum is NOT the same thing as pgavd). There was more code involved just in managing the pgavd parser than there is in all of pg_autovacuum. I think the right answer for more sophisticated configuration would involve specifying a database in which to find the pg_autovacuum table(s). > Just an idea. Mind you, I'm not so sure that we want to focus > immediately on per-table settings. I think that we want to get the > "automatic" settings working fairly well first; a lot of new DBAs > would use the per-table settings to shoot themselves in the foot. > So we need to be able to make a strong recommendation to "try the > automatic settings first." Yeah, it's probably a good idea to ensure that per-table settings involves some really conspicuous form of "foot gun" (with no kevlar socks) to discourage its use except when you _know_ what you're doing... -- let name="cbbrowne" and tld="ntlug.org" in String.concat "@" [name;tld];; http://www3.sympatico.ca/cbbrowne/nonrdbms.html Q: Can SETQ only be used with numerics? A: No, SETQ may also be used by Symbolics, and use it they do. From pgsql-performance-owner@postgresql.org Tue Nov 25 02:07:31 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4A3D4D1DA6F for ; Tue, 25 Nov 2003 06:07:29 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 54653-03 for ; Tue, 25 Nov 2003 02:06:58 -0400 (AST) Received: from smtp.pspl.co.in (unknown [202.54.11.65]) by svr1.postgresql.org (Postfix) with ESMTP id 14A63D1D695 for ; Tue, 25 Nov 2003 02:06:55 -0400 (AST) Received: (from root@localhost) by smtp.pspl.co.in (8.12.9/8.12.9) id hAP66w7v031691 for ; Tue, 25 Nov 2003 11:36:58 +0530 Received: from persistent.co.in (daithan.intranet.pspl.co.in [192.168.7.161]) (authenticated bits=0) by persistent.co.in (8.12.9/8.12.9) with ESMTP id hAP66vLp031664; Tue, 25 Nov 2003 11:36:58 +0530 Message-ID: <3FC2F17A.5050307@persistent.co.in> Date: Tue, 25 Nov 2003 11:36:50 +0530 From: Shridhar Daithankar Organization: Persistent Systems Pvt. Ltd. User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: William Yu Cc: pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? References: <3FC19C2C.4050704@myrealbox.com> In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/238 X-Sequence-Number: 4798 William Yu wrote: > This is an intriguing thought which leads me to think about a similar > solution for even a production server and that's a solid state drive for > just the WAL. What's the max disk space the WAL would ever take up? > There's quite a few 512MB/1GB/2GB solid state drives available now in > the ~$200-$500 range and if you never hit those limits... Maximum number of WAL segments at any time in 2*(number of checkpoint segments)+1 IIRC. So if you have 3 checkpoint segments, you can not have more than 7 WAL segments at any time. Give or take 1. Correct me if I am wrong.. Shridhar From pgsql-performance-owner@postgresql.org Tue Nov 25 11:52:52 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9AB24D1DC04 for ; Tue, 25 Nov 2003 14:05:46 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 57255-01 for ; Tue, 25 Nov 2003 10:05:19 -0400 (AST) Received: from mail.dsvr.co.uk (mail.dsvr.co.uk [212.69.192.9]) by svr1.postgresql.org (Postfix) with ESMTP id 53F83D1DC3E for ; Tue, 25 Nov 2003 10:05:15 -0400 (AST) Received: from dsvr.net (eth0.rt1.o1-1.tck.dsvr.net [::ffff:212.69.216.20]) by mail.dsvr.co.uk with esmtp; Tue, 25 Nov 2003 14:05:17 +0000 Message-ID: <3FC3623E.3020304@dsvr.net> Date: Tue, 25 Nov 2003 14:07:58 +0000 From: Rob Fielding User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030807 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Re: Where to start for performance problem? References: <2b1f01c3b2d6$c403ba70$0201a8c0@game> In-Reply-To: <2b1f01c3b2d6$c403ba70$0201a8c0@game> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/239 X-Sequence-Number: 4799 The problems with giving suggestions about increasing performance is that one persons increase is another persons decrease. having said that, there are a few general suggestions : Set-up some shared memory, about a tenth of your available RAM, and configure shared_memory and max_clients correctly. I've used the following formula, ripped off the net from somewhere. It's not entirely acurate, as other settings steal a little shared memory, but it works for the most part : ((1024*RAM_SIZE) - (14.2 * max_connections) - 250) / 8.2 as I say, it should get you a good value, otherwise lower it bit by bit if you have trouble starting your db. Increase effective_cache (50%-70% avail ram) and sort_mem (about 1/20th ram) and lower you random_page_cost to around 2 or less (as low as 0.3) if you have fast SCSI drives in a RAID10 set-up - this was a big speedup ;) But this might not be the answer though. The values detailed above are when tuning an already stable setup. Perhaps you need to look at your system resource usage. If you're degrading performance over time it sounds to me like you are slowly running out of memory and swap ? Generall if I take something over, I'll try and get it onto my terms. Have you tried importing the DB to a fresh installation, one where you know sensible defaults are set, so you aren't inheriting any cruft from the previous sysadmin. To be honest tho, I've never run pg so that it actually shutdown because it was running so badly - i just wouldn't think it would do that. -- Rob Fielding rob@dsvr.net www.dsvr.co.uk Development Designer Servers Ltd From pgsql-performance-owner@postgresql.org Tue Nov 25 12:04:02 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 7E571D1D8A8 for ; Tue, 25 Nov 2003 15:48:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 83828-07 for ; Tue, 25 Nov 2003 11:48:06 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 787DAD1D8DE for ; Tue, 25 Nov 2003 11:48:01 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAPFlb19027728; Tue, 25 Nov 2003 10:47:37 -0500 (EST) To: Shridhar Daithankar Cc: William Yu , pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? In-reply-to: <3FC2F17A.5050307@persistent.co.in> References: <3FC19C2C.4050704@myrealbox.com> <3FC2F17A.5050307@persistent.co.in> Comments: In-reply-to Shridhar Daithankar message dated "Tue, 25 Nov 2003 11:36:50 +0530" Date: Tue, 25 Nov 2003 10:47:37 -0500 Message-ID: <27727.1069775257@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/240 X-Sequence-Number: 4800 Shridhar Daithankar writes: > William Yu wrote: >> This is an intriguing thought which leads me to think about a similar >> solution for even a production server and that's a solid state drive for >> just the WAL. What's the max disk space the WAL would ever take up? > Maximum number of WAL segments at any time in 2*(number of checkpoint > segments)+1 IIRC. > So if you have 3 checkpoint segments, you can not have more than 7 WAL > segments at any time. Give or take 1. I don't believe that's a *hard* limit. The system tries to schedule checkpoints often enough to prevent WAL from getting bigger than that, but if you had a sufficiently big spike in update activity, it's at least theoretically possible that more than checkpoint_segments segments could be filled before the concurrently running checkpoint finishes and releases some old segments. The odds of this being a real problem are small, especially if you don't try to fit on an undersized SSD by reducing checkpoint_segments. I'd think that a 512Mb SSD would be plenty of space for ordinary update load levels ... regards, tom lane From pgsql-performance-owner@postgresql.org Tue Nov 25 12:54:08 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8DFA0D1B440 for ; Tue, 25 Nov 2003 16:54:05 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 95033-06 for ; Tue, 25 Nov 2003 12:53:33 -0400 (AST) Received: from wolff.to (wolff.to [66.93.249.74]) by svr1.postgresql.org (Postfix) with SMTP id CAD60D1B438 for ; Tue, 25 Nov 2003 12:53:33 -0400 (AST) Received: (qmail 25386 invoked by uid 500); 25 Nov 2003 16:52:32 -0000 Date: Tue, 25 Nov 2003 10:52:32 -0600 From: Bruno Wolff III To: MK Spam Cc: pgsql-performance@postgresql.org Subject: Re: Where to start for performance problem? Message-ID: <20031125165232.GA25043@wolff.to> Mail-Followup-To: MK Spam , pgsql-performance@postgresql.org References: <2b1f01c3b2d6$c403ba70$0201a8c0@game> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <2b1f01c3b2d6$c403ba70$0201a8c0@game> User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/241 X-Sequence-Number: 4801 On Mon, Nov 24, 2003 at 16:03:17 -0600, MK Spam wrote: > > The archives of this list provides many ideas for improving performance, but the problem we are having is gradually degrading performance ending in postgres shutting down. So it's not a matter of optimizing a complex query to take 5 seconds instead of 60 seconds. From what I can tell we are using the VACUUM command on a schedule but it doesn't seem to prevent the database from becoming "congested" as we refer to it. :] Anyway, the only way I know to "fix" the problem is to export (pg_dump) the db, drop the database, recreate the database and import the dump. This seems to return performance back to normal but obviously isn't a very good "solution". The slowdown and subsequent crash can take as little as 1 week for databases with a lot of data or go as long as a few weeks to a month for smaller data sets. A couple of things you might look for are index bloat and having FSM set too small for your plain vacuums. Upgrading to 7.4 may help with index bloat if that is your problem. From pgsql-performance-owner@postgresql.org Tue Nov 25 13:08:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6E737D1B448 for ; Tue, 25 Nov 2003 17:08:45 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 98193-07 for ; Tue, 25 Nov 2003 13:08:13 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id C764AD1B444 for ; Tue, 25 Nov 2003 13:07:50 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAPH7n19028468; Tue, 25 Nov 2003 12:07:49 -0500 (EST) To: MK Spam Cc: pgsql-performance@postgresql.org Subject: Re: Where to start for performance problem? In-reply-to: <20031125165232.GA25043@wolff.to> References: <2b1f01c3b2d6$c403ba70$0201a8c0@game> <20031125165232.GA25043@wolff.to> Comments: In-reply-to Bruno Wolff III message dated "Tue, 25 Nov 2003 10:52:32 -0600" Date: Tue, 25 Nov 2003 12:07:49 -0500 Message-ID: <28467.1069780069@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/242 X-Sequence-Number: 4802 MK Spam wrote: > ... the problem we are having is gradually degrading > performance ending in postgres shutting down. As someone else commented, that's not an ordinary sort of performance problem. What exactly happens when the database "shuts down"? regards, tom lane From pgsql-performance-owner@postgresql.org Tue Nov 25 14:28:24 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 831B8D1B451 for ; Tue, 25 Nov 2003 18:28:23 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 10627-07 for ; Tue, 25 Nov 2003 14:27:51 -0400 (AST) Received: from smtp-relay-7.sea.adobe.com (smtp-relay-7.adobe.com [192.150.22.7]) by svr1.postgresql.org (Postfix) with ESMTP id 404D0D1B450 for ; Tue, 25 Nov 2003 14:27:51 -0400 (AST) Received: from inner-relay-1.corp.adobe.com (inner-relay-1 [153.32.1.51]) by smtp-relay-7.sea.adobe.com (8.12.10/8.12.10) with ESMTP id hAPIRY7c003097 for ; Tue, 25 Nov 2003 10:27:34 -0800 (PST) Received: from mail-321.corp.adobe.com (mail-321 [153.32.2.46]) by inner-relay-1.corp.adobe.com (8.12.9/8.12.9) with ESMTP id hAPIRT6F005192 for ; Tue, 25 Nov 2003 10:27:29 -0800 (PST) Received: from adobe.com (c-155-236.corp.adobe.com [153.32.155.236]) by mail-321.corp.adobe.com (8.11.4/8.11.4) with ESMTP id hAPIRSA08003 for ; Tue, 25 Nov 2003 10:27:28 -0800 (PST) Message-ID: <3FC3A2A7.6060804@adobe.com> Date: Tue, 25 Nov 2003 10:42:47 -0800 From: shane hill User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5b) Gecko/20030901 Thunderbird/0.2 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance Subject: design question: general db performance References: <3FC26E96.9090101@be-a-part.de> <3FC267FB.806@bigfoot.com> <3FC27911.7070403@be-a-part.de> <3FC27071.9000001@bigfoot.com> In-Reply-To: <3FC27071.9000001@bigfoot.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/243 X-Sequence-Number: 4803 Hi folks, Disclaimer: I am relatively new to RDBMSs, so please do not laugh at me too loudly, you can laugh, just not too loudly and please do not point. :) I am working on an Automated Installer Testing System for Adobe Systems and I am doing a DB redesign of the current postgres db: 1. We are testing a matrix of over 900 Acrobat installer configurations and we are tracking every file and registry entry that is affected by an installation. 2. a single file or registry entry that is affected by any test is stored in the db as a record. 3. a typical record is about 12 columns of string data. the data is all information about a file (mac or windows) or windows registry entry [ file or regkey name, file size, modification date, checksum, permissions, owner, group, and in the case of a mac, we are getting all the hfs atts as well]. 4. A typical test produces anywhere from 2000 - 5000 records. Our db is getting to be a respectable size (about 10GB right now) and is growing slower and slower. I have been charged with making it faster and with a smaller footprint while retaining all of the current functionality. here is one of my ideas. Please tell me if I am crazy: The strings that we are storing (mentioned in 3 above) are extremely repetitive. for example, there are a limited number of permissions for the files in the acrobat installer and we are storing this information over and over again in the tables. The same goes for filenames, registry key names and almost all of the data we are storing. So it seems to me that to create a smaller and faster database we could assign an integer to each string and just store the integer representation of the string rather than the string itself. Then we would just store the strings in a separate table one time and do join queries against the tables that are holding the strings and the main data tables. for example, a table that would hold unique permissions strings would look like table: perms_strs string | id --------------------- 'drwxr-xr-x' | 1 '-rw-------' | 2 'drwxrwxr-x' | 3 '-rw-r--r--' | 4 then in my data I would just store 1,2,3 or 4 instead of the whole permissions string. it seems to me that we would save lots of space and over time not see the same performance degradation. anyways, please tell me if this makes sense and make any other suggestions that you can think of. I am just now starting this analysis so I cannot give specifics as to where we are seeing poor performance just yet. just tell me if my concepts are correct. thanks for your time and for suffering this email. chao, -Shane From pgsql-performance-owner@postgresql.org Tue Nov 25 14:59:03 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id D1597D1B45E for ; Tue, 25 Nov 2003 18:59:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 17122-05 for ; Tue, 25 Nov 2003 14:58:29 -0400 (AST) Received: from indygecko.com (h24-71-76-41.ok.shawcable.net [24.71.76.41]) by svr1.postgresql.org (Postfix) with ESMTP id 4B8DFD1B447 for ; Tue, 25 Nov 2003 14:58:29 -0400 (AST) Received: from [192.168.10.10] ([::ffff:192.168.10.10]) by indygecko.com with esmtp; Tue, 25 Nov 2003 10:58:28 -0800 Subject: Re: design question: general db performance From: Jord Tanner To: shane hill Cc: pgsql-performance In-Reply-To: <3FC3A2A7.6060804@adobe.com> References: <3FC26E96.9090101@be-a-part.de> <3FC267FB.806@bigfoot.com> <3FC27911.7070403@be-a-part.de> <3FC27071.9000001@bigfoot.com> <3FC3A2A7.6060804@adobe.com> Organization: Message-Id: <1069786707.4242.66.camel@gecko> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.2.2 (1.2.2-5) Date: 25 Nov 2003 10:58:28 -0800 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/244 X-Sequence-Number: 4804 [small chuckle] By George, I think he's got it! You are on the right track. Have a look at this link on database normalization for more info: http://databases.about.com/library/weekly/aa080501a.htm On Tue, 2003-11-25 at 10:42, shane hill wrote: > Hi folks, > > Disclaimer: I am relatively new to RDBMSs, so please do not laugh at me > too loudly, you can laugh, just not too loudly and please do not point. :) > [snip] -- Jord Tanner From pgsql-performance-owner@postgresql.org Fri Nov 28 14:15:34 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CBDB9D1B46F for ; Tue, 25 Nov 2003 19:05:05 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 16929-08 for ; Tue, 25 Nov 2003 15:04:33 -0400 (AST) Received: from node-4024d1a2.mdw.onnet.us.uu.net (node-40241d3a.mdw.onnet.us.uu.net [64.36.29.58]) by svr1.postgresql.org (Postfix) with ESMTP id B3A1FD1B45F for ; Tue, 25 Nov 2003 15:04:33 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by node-4024d1a2.mdw.onnet.us.uu.net (Postfix) with ESMTP id 2BA7637D0B for ; Tue, 25 Nov 2003 12:59:54 -0600 (CST) Subject: Re: Maximum Possible Insert Performance? From: Suchandra Thapa To: pgsql-performance@postgresql.org In-Reply-To: <87brr11c44.fsf@stark.dyndns.tv> References: <200311240925.58070.josh@agliodbs.com> <200311241004.37969.josh@agliodbs.com> <87brr11c44.fsf@stark.dyndns.tv> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-by9rSZ6nTuoexoHmUfLz" Organization: Message-Id: <1069786793.19104.33.camel@hepcat> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.2.2 (1.2.2-5) Date: 25 Nov 2003 12:59:54 -0600 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/298 X-Sequence-Number: 4858 --=-by9rSZ6nTuoexoHmUfLz Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Mon, 2003-11-24 at 19:16, Greg Stark wrote: > William Yu writes: >=20 > > > You're right, though, mirroring a solid state drive is pretty pointle= ss; if > > > power fails, both mirrors are dead. > >=20 > > Actually no. Solid state memory is non-volatile. They retain data even = without > > power. >=20 > Note that flash ram only has a finite number of write cycles before it fa= ils. >=20 > On the other hand that might not be so bad for WAL which writes sequentia= lly, > you can easily calculate how close you are to the maximum. For things like > heap storage or swap it's awful as you can get hot spots that get written= to > thousands of times before the rest of the space is used. I could be wrong, but I was under the impression that most of the newer flash disks tended to spread writes out over the drive so that hotspots are minimized.=20=20 --=20 Suchandra Thapa --=-by9rSZ6nTuoexoHmUfLz Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQA/w6ap6nShCjt5AZIRAii6AKCfCeSjosYKgxEJi628ftZwpqO7JACeMGUl fr3exMlo0k2CFdP57BGevPs= =S6Ro -----END PGP SIGNATURE----- --=-by9rSZ6nTuoexoHmUfLz-- From pgsql-performance-owner@postgresql.org Tue Nov 25 15:23:54 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 40014D1B45F for ; Tue, 25 Nov 2003 19:23:50 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 22406-01 for ; Tue, 25 Nov 2003 15:23:18 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id DB17BD1B434 for ; Tue, 25 Nov 2003 15:23:17 -0400 (AST) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3977127; Tue, 25 Nov 2003 11:23:50 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: shane hill , pgsql-performance Subject: Re: design question: general db performance Date: Tue, 25 Nov 2003 11:12:47 -0800 User-Agent: KMail/1.4.3 References: <3FC26E96.9090101@be-a-part.de> <3FC27071.9000001@bigfoot.com> <3FC3A2A7.6060804@adobe.com> In-Reply-To: <3FC3A2A7.6060804@adobe.com> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200311251112.47473.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/245 X-Sequence-Number: 4805 Shane, > Disclaimer: I am relatively new to RDBMSs, so please do not laugh at me= =20 > too loudly, you can laugh, just not too loudly and please do not point. = :) Hey, we all started somewhere. Nobody was born knowing databases. Except= =20 maybe Neil Conway. > I am working on an Automated Installer Testing System for Adobe Systems= =20 > and I am doing a DB redesign of the current postgres db: Cool! We're going to want to talk to you about a case study later, if yo= u=20 can get your boss to authorize it .... > Our db is getting to be a respectable size (about 10GB right now) and is= =20 > growing slower and slower.=20 Slower and slower? Hmmm ... what's your VACUUM. ANALYZE & REINDEX schedul= e?=20=20 What PG version? What are your postgresql.conf settings? Progressive=20 performance loss may indicate a problem with one or more of these things ... > then in my data I would just store 1,2,3 or 4 instead of the whole=20 > permissions string. >=20 > it seems to me that we would save lots of space and over time not see=20 > the same performance degradation. Yes, this is a good idea. Abstracting other repetitive data is good too.= =20=20 Also keep in mind that the permissions themselves can be represented as oct= al=20 numbers instead of strings, which takes less space. --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Nov 25 16:21:13 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 245A3D1B47D for ; Tue, 25 Nov 2003 20:20:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 24158-09 for ; Tue, 25 Nov 2003 16:20:09 -0400 (AST) Received: from pns.mm.eutelsat.org (pns.mm.eutelsat.org [194.214.173.227]) by svr1.postgresql.org (Postfix) with ESMTP id 3B905D1B436 for ; Tue, 25 Nov 2003 16:20:08 -0400 (AST) Received: from nts-03.mm.eutelsat.org (localhost [127.0.0.1]) by pns.mm.eutelsat.org (8.11.6/linuxconf) with ESMTP id hAPKNEA11370; Tue, 25 Nov 2003 21:23:14 +0100 Received: from bigfoot.com (accesspoint.mm.eutelsat.org [194.214.173.4]) by nts-03.mm.eutelsat.org (8.11.6/linuxconf) with ESMTP id hAPK8vR32160; Tue, 25 Nov 2003 21:08:57 +0100 Message-ID: <3FC3AB3D.4080708@bigfoot.com> Date: Tue, 25 Nov 2003 20:19:25 +0100 From: Gaetano Mendola User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: "pgsql-performance@postgresql.org" Cc: Torsten Schulz Subject: Re: [Fwd: Re: Optimize] References: <3FC3AFD9.300@be-a-part.de> In-Reply-To: <3FC3AFD9.300@be-a-part.de> X-Enigmail-Version: 0.81.7.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/252 X-Sequence-Number: 4812 Torsten Schulz wrote: > Hi, > >> You can see doing select * from pg_stat_activity the >> queries that are currently running on your server, and >> do a explain analize on it to see which one is the >> bottleneck. If you are running the 7.4 you can see on >> the log the total ammount for each query. > > > > with this query I see how much queries running, but the field > current_query are free, so i can't see which queries are very slow. You must perform that query with permission of super_user. Regards Gaetano Mendola From pgsql-performance-owner@postgresql.org Tue Nov 25 15:24:07 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 75203D1B460 for ; Tue, 25 Nov 2003 19:24:05 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 21749-06 for ; Tue, 25 Nov 2003 15:23:34 -0400 (AST) Received: from jefftrout.com (h00a0cc4084e5.ne.client2.attbi.com [24.128.241.68]) by svr1.postgresql.org (Postfix) with SMTP id E61A7D1B453 for ; Tue, 25 Nov 2003 15:23:33 -0400 (AST) Received: (qmail 89861 invoked from network); 25 Nov 2003 19:23:39 -0000 Received: from localhost (HELO squeegit) (threshar@127.0.0.1) by localhost with SMTP; 25 Nov 2003 19:23:39 -0000 Date: Tue, 25 Nov 2003 14:23:18 -0500 From: Jeff To: shane hill Cc: pgsql-performance@postgresql.org Subject: Re: design question: general db performance Message-Id: <20031125142318.4ff8bdb2.threshar@torgo.978.org> In-Reply-To: <3FC3A2A7.6060804@adobe.com> References: <3FC26E96.9090101@be-a-part.de> <3FC267FB.806@bigfoot.com> <3FC27911.7070403@be-a-part.de> <3FC27071.9000001@bigfoot.com> <3FC3A2A7.6060804@adobe.com> X-Mailer: Sylpheed version 0.9.7 (GTK+ 1.2.10; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/246 X-Sequence-Number: 4806 On Tue, 25 Nov 2003 10:42:47 -0800 shane hill wrote: > Our db is getting to be a respectable size (about 10GB right now) and > is growing slower and slower. I have been charged with making it > faster and with a smaller footprint while retaining all of the current > functionality. here is one of my ideas. Please tell me if I am > crazy: > What exactly is it getting slower doing? Have you run through the usual gamut of things to check - shared buffers, vacuum analyzig, etc. etc. What ver of PG? What OS? Can you post any schema/queries? Normalizing can help. But I don't think it is going to be a magical bullet that will make the DB instantly fast. It will reduce the size of it though. -- Jeff Trout http://www.jefftrout.com/ http://www.stuarthamm.net/ From pgsql-performance-owner@postgresql.org Tue Nov 25 15:28:40 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0AEFDD1B47A for ; Tue, 25 Nov 2003 19:28:32 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 21190-09 for ; Tue, 25 Nov 2003 15:28:00 -0400 (AST) Received: from anchor-post-32.mail.demon.net (anchor-post-32.mail.demon.net [194.217.242.90]) by svr1.postgresql.org (Postfix) with ESMTP id 6381ED1B434 for ; Tue, 25 Nov 2003 15:27:44 -0400 (AST) Received: from mwynhau.demon.co.uk ([193.237.186.96] helo=mainbox.archonet.com) by anchor-post-32.mail.demon.net with esmtp (Exim 3.35 #1) id 1AOiqd-0000Hj-0W; Tue, 25 Nov 2003 19:27:43 +0000 Received: by mainbox.archonet.com (Postfix, from userid 529) id D42AE17DE4; Tue, 25 Nov 2003 19:27:41 +0000 (GMT) Received: from client17.archonet.com (client17.archonet.com [192.168.1.17]) by mainbox.archonet.com (Postfix) with ESMTP id 190DB179A4; Tue, 25 Nov 2003 19:27:40 +0000 (GMT) From: Richard Huxton To: shane hill , pgsql-performance Subject: Re: design question: general db performance Date: Tue, 25 Nov 2003 19:27:39 +0000 User-Agent: KMail/1.5 References: <3FC26E96.9090101@be-a-part.de> <3FC27071.9000001@bigfoot.com> <3FC3A2A7.6060804@adobe.com> In-Reply-To: <3FC3A2A7.6060804@adobe.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311251927.39671.dev@archonet.com> X-Bogosity: No, tests=bogofilter, spamicity=0.000000, version=0.15.3 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/247 X-Sequence-Number: 4807 On Tuesday 25 November 2003 18:42, shane hill wrote: > > Our db is getting to be a respectable size (about 10GB right now) and is > growing slower and slower. I have been charged with making it faster and > with a smaller footprint while retaining all of the current > functionality. here is one of my ideas. Please tell me if I am crazy: Your idea of using an integer makes sense - that's how it is stored on unix anyway. Are you familiar with VACUUM/VACUUM FULL/REINDEX and when you should use them? If not, that's a good place to start. Try a VACUUM FULL on frequently updated tables and see if that reduces your disk size. You'll probably want to check the performance notes too: http://www.varlena.com/varlena/GeneralBits/Tidbits/index.php -- Richard Huxton Archonet Ltd From pgsql-performance-owner@postgresql.org Tue Nov 25 15:39:59 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 5D8B0D1B460 for ; Tue, 25 Nov 2003 19:39:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 23954-04 for ; Tue, 25 Nov 2003 15:39:22 -0400 (AST) Received: from natsmtp01.rzone.de (natsmtp01.rzone.de [81.169.145.166]) by svr1.postgresql.org (Postfix) with ESMTP id 1D146D1B439 for ; Tue, 25 Nov 2003 15:39:22 -0400 (AST) Received: from be-a-part.de (pD9E81E1A.dip.t-dialin.net [217.232.30.26]) by post.webmailer.de (8.12.10/8.12.10) with ESMTP id hAPJd5SV002529 for ; Tue, 25 Nov 2003 20:39:06 +0100 (MET) Message-ID: <3FC3AFD9.300@be-a-part.de> Date: Tue, 25 Nov 2003 20:39:05 +0100 From: Torsten Schulz User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5b) Gecko/20030827 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: [Fwd: Re: Optimize] Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.8 tagged_above=0.0 required=5.0 tests=RCVD_IN_NJABL, USER_AGENT_MOZILLA_UA X-Spam-Level: X-Archive-Number: 200311/248 X-Sequence-Number: 4808 Hi, > You can see doing select * from pg_stat_activity the > queries that are currently running on your server, and > do a explain analize on it to see which one is the > bottleneck. If you are running the 7.4 you can see on > the log the total ammount for each query. with this query I see how much queries running, but the field current_query are free, so i can't see which queries are very slow. Greetings Torsten From pgsql-performance-owner@postgresql.org Tue Nov 25 15:47:06 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 17E5FD1B461 for ; Tue, 25 Nov 2003 19:47:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 25699-01 for ; Tue, 25 Nov 2003 15:46:28 -0400 (AST) Received: from jhuml2.jhmi.edu (jhuml2.jhmi.edu [162.129.234.21]) by svr1.postgresql.org (Postfix) with ESMTP id 97E0ED1B454 for ; Tue, 25 Nov 2003 15:46:27 -0400 (AST) Received: from jhuml2.jhmi.edu (jhuml2.jhmi.edu [162.129.234.21]) by jhuml2.jhmi.edu (PMDF V6.2-X17 #30840) with SMTP id <0HOX005BYAR30W@jhuml2.jhmi.edu> for pgsql-performance@postgresql.org; Tue, 25 Nov 2003 14:45:44 -0500 (EST) Received: from jhuml2.jhmi.edu ([162.129.234.21]) by jhuml2.jhmi.edu (SAVSMTP 3.1.0.29) with SMTP id M2003112514454426245 for ; Tue, 25 Nov 2003 14:45:44 -0500 Received: from jhmimail.jhmi.edu (jhem2.jhmi.edu [162.129.8.23]) by jhuml2.jhmi.edu (PMDF V6.2-X17 #30840) with ESMTP id <0HOX0039TAW7XU@jhuml2.jhmi.edu> for pgsql-performance@postgresql.org; Tue, 25 Nov 2003 14:45:44 -0500 (EST) Received: from [162.129.178.60] by jhmimail.jhmi.edu (mshttpd); Tue, 25 Nov 2003 19:48:49 +0000 (GMT) Date: Tue, 25 Nov 2003 19:48:49 +0000 (GMT) From: LIANHE SHAO Subject: why index scan not working when using 'like'? To: pgsql-performance@postgresql.org Message-id: <3d83693d9fe0.3d9fe03d8369@jhmimail.jhmi.edu> MIME-version: 1.0 X-Mailer: iPlanet Messenger Express 5.2 HotFix 1.17 (built Jun 23 2003) Content-type: text/plain; charset=us-ascii Content-language: en Content-transfer-encoding: 7bit Content-disposition: inline X-Accept-Language: en X-Virus-Scanned: by amavisd-new at postgresql.org X-Spam-Status: No, hits=0.0 tagged_above=0.0 required=5.0 tests= X-Spam-Level: X-Archive-Number: 200311/249 X-Sequence-Number: 4809 Hi all, I want to use index on the gene_symbol column in my query and gene_symbol is indexed. but when I use lower (gene_symbol) like lower('%mif%'), the index is not used. While when I change to lower(gene_symbol) = lower('mif'), the index is used and index scan works, but this is not what I like. I want all the gene_symbols containing substring 'mif' are pulled out, and not necessarily exactly match. could anybody give me some hints how to deal with this. If I do not used index, it take too long for the query. PGA> explain select distinct probeset_id, chip, gene_symbol, title, sequence_description, pfam from affy_array_annotation where lower(gene_symbol) like upper('%mif%'); QUERY PLAN ----------------------------------------------------------------------------------------- Unique (cost=29576.44..29591.44 rows=86 width=265) -> Sort (cost=29576.44..29578.59 rows=857 width=265) Sort Key: probeset_id, chip, gene_symbol, title, sequence_description, pfam -> Seq Scan on affy_array_annotation (cost=0.00..29534.70 rows=857 width=265) Filter: (lower((gene_symbol)::text) ~~ 'MIF%'::text) (5 rows) PGA=> explain select distinct probeset_id, chip, gene_symbol, title, sequence_description, pfam from affy_array_annotation where lower(gene_symbol) = upper('%mif%'); QUERY PLAN --------------------------------------------------------------------------------------------------------------------- Unique (cost=3433.44..3448.44 rows=86 width=265) -> Sort (cost=3433.44..3435.58 rows=857 width=265) Sort Key: probeset_id, chip, gene_symbol, title, sequence_description, pfam -> Index Scan using gene_symbol_idx_fun1 on affy_array_annotation (cost=0.00..3391.70 rows=857 width=265) Index Cond: (lower((gene_symbol)::text) = '%MIF%'::text) (5 rows) Regards, William From pgsql-performance-owner@postgresql.org Tue Nov 25 16:02:29 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C1AD2D1B437 for ; Tue, 25 Nov 2003 20:02:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 27088-01 for ; Tue, 25 Nov 2003 16:01:56 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id C03E6D1B431 for ; Tue, 25 Nov 2003 16:01:55 -0400 (AST) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3977322; Tue, 25 Nov 2003 12:02:44 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: LIANHE SHAO , pgsql-performance@postgresql.org Subject: Re: why index scan not working when using 'like'? Date: Tue, 25 Nov 2003 11:51:40 -0800 User-Agent: KMail/1.4.3 References: <3d83693d9fe0.3d9fe03d8369@jhmimail.jhmi.edu> In-Reply-To: <3d83693d9fe0.3d9fe03d8369@jhmimail.jhmi.edu> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200311251151.40781.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/251 X-Sequence-Number: 4811 Lianhe, > I want to use index on the gene_symbol column in my > query and gene_symbol is indexed. but when I use > lower (gene_symbol) like lower('%mif%'), the index > is not used. While when I change to > lower(gene_symbol) =3D lower('mif'), the index is used > and index scan works, but this is not what I like. I > want all the gene_symbols containing substring > 'mif' are pulled out, and not necessarily exactly match. LIKE '%mif%' is what's called an "unanchored text search" and it cannot use= an=20 index. The database *has* to scan the full text looking for the substring= .=20=20 This is true of all database platforms I know of. In regular text fields containing words, your problem is solvable with full= =20 text indexing (FTI). Unfortunately, FTI is not designed for arbitrary=20 non-language strings. It could be adapted, but would require a lot of=20 hacking. So you will need to find a way to restructure you data to avoid needing=20 unanchored text searches. One way would be to break down the gene_symbol= =20 field into its smallest atomic components and store those in an indexed chi= ld=20 table. Or if you're searching on the same values all the time, you can=20 create a partial index. --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Nov 25 15:57:09 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A1822D1B47D for ; Tue, 25 Nov 2003 19:57:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 26341-02 for ; Tue, 25 Nov 2003 15:56:28 -0400 (AST) Received: from rlx13.zapatec.com (66-117-144-213.zapatec.lmi.net [66.117.144.213]) by svr1.postgresql.org (Postfix) with ESMTP id 72EF0D1B43C for ; Tue, 25 Nov 2003 15:56:28 -0400 (AST) Received: from rlx11.zapatec.com (rlx11.pr.zapatec.com [192.168.1.132]) by rlx13.zapatec.com (Postfix) with ESMTP id D6D19A941 for ; Tue, 25 Nov 2003 11:56:14 -0800 (PST) Received: (from dror@localhost) by rlx11.zapatec.com (8.12.3/8.12.3/Submit) id hAPJuDaH040935 for pgsql-performance@postgresql.org; Tue, 25 Nov 2003 11:56:13 -0800 (PST) (envelope-from dror) Date: Tue, 25 Nov 2003 11:56:13 -0800 From: Dror Matalon To: pgsql-performance@postgresql.org Subject: Re: why index scan not working when using 'like'? Message-ID: <20031125195613.GB30893@rlx11.zapatec.com> References: <3d83693d9fe0.3d9fe03d8369@jhmimail.jhmi.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3d83693d9fe0.3d9fe03d8369@jhmimail.jhmi.edu> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/250 X-Sequence-Number: 4810 Hi, Searches with like or regexes often can't use the index. Think of the index as a sorted list of your items. It's easy to find an item when you know it starts with mif so ('mif%' should use the index). But when you use a 'like' that starts with '%' the index is useless and the search needs to do a sequential scan. Regards, Dror On Tue, Nov 25, 2003 at 07:48:49PM +0000, LIANHE SHAO wrote: > Hi all, > > I want to use index on the gene_symbol column in my > query and gene_symbol is indexed. but when I use > lower (gene_symbol) like lower('%mif%'), the index > is not used. While when I change to > lower(gene_symbol) = lower('mif'), the index is used > and index scan works, but this is not what I like. I > want all the gene_symbols containing substring > 'mif' are pulled out, and not necessarily exactly match. > > could anybody give me some hints how to deal with > this. If I do not used index, it take too long for > the query. > > > PGA> explain select distinct probeset_id, chip, > gene_symbol, title, sequence_description, pfam from > affy_array_annotation where lower(gene_symbol) like > upper('%mif%'); > QUERY PLAN > ----------------------------------------------------------------------------------------- > Unique (cost=29576.44..29591.44 rows=86 width=265) > -> Sort (cost=29576.44..29578.59 rows=857 > width=265) > Sort Key: probeset_id, chip, gene_symbol, > title, sequence_description, pfam > -> Seq Scan on affy_array_annotation > (cost=0.00..29534.70 rows=857 width=265) > Filter: (lower((gene_symbol)::text) > ~~ 'MIF%'::text) > (5 rows) > > > PGA=> explain select distinct probeset_id, chip, > gene_symbol, title, sequence_description, pfam from > affy_array_annotation where lower(gene_symbol) = > upper('%mif%'); > > QUERY PLAN > --------------------------------------------------------------------------------------------------------------------- > Unique (cost=3433.44..3448.44 rows=86 width=265) > -> Sort (cost=3433.44..3435.58 rows=857 width=265) > Sort Key: probeset_id, chip, gene_symbol, > title, sequence_description, pfam > -> Index Scan using gene_symbol_idx_fun1 > on affy_array_annotation (cost=0.00..3391.70 > rows=857 width=265) > Index Cond: > (lower((gene_symbol)::text) = '%MIF%'::text) > (5 rows) > > > > > > Regards, > William > > > ---------------------------(end of broadcast)--------------------------- > TIP 8: explain analyze is your friend -- Dror Matalon Zapatec Inc 1700 MLK Way Berkeley, CA 94709 http://www.fastbuzz.com http://www.zapatec.com From pgsql-performance-owner@postgresql.org Tue Nov 25 17:29:43 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 6F2B8D1B44F for ; Tue, 25 Nov 2003 21:29:37 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 37579-09 for ; Tue, 25 Nov 2003 17:29:06 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 671F0D1B474 for ; Tue, 25 Nov 2003 17:29:05 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAPLT419008755; Tue, 25 Nov 2003 16:29:04 -0500 (EST) To: josh@agliodbs.com Cc: LIANHE SHAO , pgsql-performance@postgresql.org Subject: Re: why index scan not working when using 'like'? In-reply-to: <200311251151.40781.josh@agliodbs.com> References: <3d83693d9fe0.3d9fe03d8369@jhmimail.jhmi.edu> <200311251151.40781.josh@agliodbs.com> Comments: In-reply-to Josh Berkus message dated "Tue, 25 Nov 2003 11:51:40 -0800" Date: Tue, 25 Nov 2003 16:29:04 -0500 Message-ID: <8754.1069795744@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/253 X-Sequence-Number: 4813 Josh Berkus writes: > In regular text fields containing words, your problem is solvable with full > text indexing (FTI). Unfortunately, FTI is not designed for arbitrary > non-language strings. It could be adapted, but would require a lot of > hacking. I'm not sure why you say that FTI isn't a usable solution. As long as the gene symbols are separated by whitespace or some other non-letters (eg, "foo mif bar" not "foomifbar"), I'd think FTI would work. regards, tom lane From pgsql-performance-owner@postgresql.org Tue Nov 25 17:58:30 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 7E28AD1B48D for ; Tue, 25 Nov 2003 21:58:28 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 42250-10 for ; Tue, 25 Nov 2003 17:57:58 -0400 (AST) Received: from smtp2.mail.be.easynet.net (smtp2.mail.be.easynet.net [212.100.160.76]) by svr1.postgresql.org (Postfix) with ESMTP id 84AFAD1B45F for ; Tue, 25 Nov 2003 17:57:57 -0400 (AST) Received: from 213-193-177-64.adsl.easynet.be ([213.193.177.64]) by smtp2.mail.be.easynet.net with esmtp (Exim 4.22) id 1AOlBs-0004mZ-9Y for pgsql-performance@postgresql.org; Tue, 25 Nov 2003 22:57:48 +0100 From: Stefan Champailler Reply-To: schampailler@easynet.be To: pgsql-performance@postgresql.org Subject: Impossibly slow DELETEs Date: Tue, 25 Nov 2003 22:56:56 +0100 User-Agent: KMail/1.5.3 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311252256.56433.schampailler@easynet.be> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/254 X-Sequence-Number: 4814 Dear You all, (please tell me if this has already been discussed, I was unable to find any convincing information) I'm developing a small application, tied to a PG 7.4 beta 5 (i didn't upgrade). The DB i use is roughly 20 tales each of them containing at most 30 records (I'm still in development). I can provide a whole dump if necessary. I access the DB throug IODBC (Suse Linux 8.1), through PHP. The machine everything runs on is 512M of Ram, 2.5GHz speed. So I assume it should be blazingly fast. So here's my trouble : some DELETE statement take up to 1 minute to complete (but not always, sometimes it's fast, sometimes it's that slow). Here's a typical one : DELETE FROM response_bool WHERE response_id = '125' The response_bool table has no foreing key and no index on response_id column. No foreign key reference the response_bool table. There are 6 rows in the table (given that size, I assumed that an index was not necessary). So 1 minute to complete look like I did something REALLY bad. It is my feeling that doing the same query with psql works without problem, but I can't be sure. The rest of my queries (inserts, updates) just work fine and pretty fast. Can someone help me or point me to a place where I can find help ? I didn't do any in deep debugging though. thx, stF From pgsql-performance-owner@postgresql.org Tue Nov 25 18:06:43 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 9A6BED1B474 for ; Tue, 25 Nov 2003 22:06:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 45480-02 for ; Tue, 25 Nov 2003 18:06:10 -0400 (AST) Received: from natsmtp01.rzone.de (natsmtp01.rzone.de [81.169.145.166]) by svr1.postgresql.org (Postfix) with ESMTP id 1B3E1D1B45F for ; Tue, 25 Nov 2003 18:06:09 -0400 (AST) Received: from be-a-part.de (pD9E81E1A.dip.t-dialin.net [217.232.30.26]) by post.webmailer.de (8.12.10/8.12.10) with ESMTP id hAPM6AmC002378 for ; Tue, 25 Nov 2003 23:06:10 +0100 (MET) Message-ID: <3FC3D252.1000204@be-a-part.de> Date: Tue, 25 Nov 2003 23:06:10 +0100 From: Torsten Schulz User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5b) Gecko/20030827 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Re: [Fwd: Re: Optimize] References: <3FC3AFD9.300@be-a-part.de> <3FC3AB3D.4080708@bigfoot.com> In-Reply-To: <3FC3AB3D.4080708@bigfoot.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/255 X-Sequence-Number: 4815 Gaetano Mendola wrote: > Torsten Schulz wrote: > >> Hi, >> >>> You can see doing select * from pg_stat_activity the >>> queries that are currently running on your server, and >>> do a explain analize on it to see which one is the >>> bottleneck. If you are running the 7.4 you can see on >>> the log the total ammount for each query. >> >> >> >> >> with this query I see how much queries running, but the field >> current_query are free, so i can't see which queries are very slow. > > > You must perform that query with permission of super_user. > I've made it in root-account with psql -U postgres - but i can't see the query Regards Torsten Schulz From pgsql-performance-owner@postgresql.org Tue Nov 25 18:08:03 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 01A1AD1B460 for ; Tue, 25 Nov 2003 22:07:51 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 43550-08 for ; Tue, 25 Nov 2003 18:07:20 -0400 (AST) Received: from natsmtp00.webmailer.de (natsmtp00.rzone.de [81.169.145.165]) by svr1.postgresql.org (Postfix) with ESMTP id A329ED1B43F for ; Tue, 25 Nov 2003 18:07:18 -0400 (AST) Received: from be-a-part.de (pD9E81E1A.dip.t-dialin.net [217.232.30.26]) by post.webmailer.de (8.12.10/8.12.10) with ESMTP id hAPM7Jkg012396 for ; Tue, 25 Nov 2003 23:07:19 +0100 (MET) Message-ID: <3FC3D296.9040404@be-a-part.de> Date: Tue, 25 Nov 2003 23:07:18 +0100 From: Torsten Schulz User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5b) Gecko/20030827 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Re: Optimize Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/256 X-Sequence-Number: 4816 Chester Kustarz wrote: > On Mon, 24 Nov 2003, Torsten Schulz wrote: > > >> shared_buffers = 5000 # 2*max_connections, min 16 >> > > > that looks pretty small. that would only be 40MBytes (8k/page * > 5000pages). > > http://www.varlena.com/GeneralBits/Tidbits/perf.html > > > Ok, thats it. I've set it to 51200, now it seems to be very fast. Thank you! -------- Original Message -------- Subject: Re: [PERFORM] Optimize Date: Tue, 25 Nov 2003 23:04:06 +0100 From: Torsten Schulz To: Chester Kustarz References: Chester Kustarz wrote: >On Mon, 24 Nov 2003, Torsten Schulz wrote: > > >>shared_buffers = 5000 # 2*max_connections, min 16 >> >> > >that looks pretty small. that would only be 40MBytes (8k/page * 5000pages). > >http://www.varlena.com/GeneralBits/Tidbits/perf.html > > > Ok, thats it. I've set it to 51200, now it seems to be very fast. Thank you! From pgsql-performance-owner@postgresql.org Tue Nov 25 18:30:28 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4E953D1B456 for ; Tue, 25 Nov 2003 22:30:21 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 44835-07 for ; Tue, 25 Nov 2003 18:29:51 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 6F08CD1B487 for ; Tue, 25 Nov 2003 18:29:50 -0400 (AST) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3977968 for pgsql-performance@postgresql.org; Tue, 25 Nov 2003 14:30:39 -0800 Content-Type: text/plain; charset="us-ascii" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: pgsql-performance@postgresql.org Subject: Wierd context-switching issue on Xeon Date: Tue, 25 Nov 2003 14:19:36 -0800 User-Agent: KMail/1.4.3 MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200311251419.36771.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/258 X-Sequence-Number: 4818 Folks, We're seeing some odd issues with hyperthreading-capable Xeons, whether or = not=20 hyperthreading is enabled. Basically, when a small number of really-heavy= =20 duty queries hit the system and push all of the CPUs to more than 70% used= =20 (about 1/2 user & 1/2 kernel), the system goes to 100,000+ context switcthe= s=20 per second and performance degrades.=20=20 I know that there's other Xeon users on this list ... has anyone else seen= =20 anything like that? The machines are Dells running Red Hat 7.3. --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Nov 25 18:27:46 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 03D7AD1B46D for ; Tue, 25 Nov 2003 22:27:37 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 43363-09 for ; Tue, 25 Nov 2003 18:27:07 -0400 (AST) Received: from mta1.sucs.soton.ac.uk (mta1.sucs.soton.ac.uk [152.78.128.140]) by svr1.postgresql.org (Postfix) with ESMTP id 155BCD1B430 for ; Tue, 25 Nov 2003 18:27:05 -0400 (AST) Received: from russell (ip-002.area-012.loc-245.halls.soton.ac.uk [10.245.12.2]) (authenticated bits=0) by mta1.sucs.soton.ac.uk (8.12.10/8.12.10) with ESMTP id hAPMQi3A008276; Tue, 25 Nov 2003 22:26:44 GMT From: "Russell Garrett" To: "Torsten Schulz" , Subject: Re: [Fwd: Re: Optimize] Date: Tue, 25 Nov 2003 22:27:34 -0000 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0) In-Reply-To: <3FC3D252.1000204@be-a-part.de> X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal X-ISS-MailScanner: Believed to be clean X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/257 X-Sequence-Number: 4817 >>> with this query I see how much queries running, but the field >>> current_query are free, so i can't see which queries are very slow. >> >> >> You must perform that query with permission of super_user. >> > I've made it in root-account with psql -U postgres - but i can't see > the query You must have these lines in your postgresql.conf for the query stats to be collected: stats_start_collector = true stats_command_string = true -------------------------------------------------------------------- Russ Garrett russ@last.fm http://last.fm From pgsql-performance-owner@postgresql.org Tue Nov 25 19:48:30 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A649FD1B441 for ; Tue, 25 Nov 2003 23:48:27 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 56083-04 for ; Tue, 25 Nov 2003 19:47:57 -0400 (AST) Received: from davinci.ethosmedia.com (server228.ethosmedia.com [209.128.84.228]) by svr1.postgresql.org (Postfix) with ESMTP id 7CF28D1B43B for ; Tue, 25 Nov 2003 19:47:55 -0400 (AST) Received: from [66.219.92.2] (HELO temoku) by davinci.ethosmedia.com (CommuniGate Pro SMTP 4.0.2) with ESMTP id 3978264; Tue, 25 Nov 2003 15:48:45 -0800 Content-Type: text/plain; charset="iso-8859-1" From: Josh Berkus Reply-To: josh@agliodbs.com Organization: Aglio Database Solutions To: Tom Lane Subject: Re: Wierd context-switching issue on Xeon Date: Tue, 25 Nov 2003 15:37:42 -0800 User-Agent: KMail/1.4.3 Cc: pgsql-performance@postgresql.org References: <200311251419.36771.josh@agliodbs.com> <9470.1069803647@sss.pgh.pa.us> In-Reply-To: <9470.1069803647@sss.pgh.pa.us> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Message-Id: <200311251537.42527.josh@agliodbs.com> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/261 X-Sequence-Number: 4821 Tom, > Strictly a WAG ... but what this sounds like to me is disastrously bad > behavior of the spinlock code under heavy contention. We thought we'd > fixed the spinlock code for SMP machines awhile ago, but maybe > hyperthreading opens some new vistas for misbehavior ... Yeah, I thought of that based on the discussion on -Hackers. But we tried= =20 turning off hyperthreading, with no change in behavior. > If you can't try 7.4, or want to gather more data first, it would be > good to try to confirm or disprove the theory that the context switches > are coming from spinlock delays. If they are, they'd be coming from the > select() calls in s_lock() in s_lock.c. Can you strace or something to > see what kernel calls the context switches occur on? Might be worth it ... will suggest that. Will also try 7.4. --=20 -Josh Berkus Aglio Database Solutions San Francisco From pgsql-performance-owner@postgresql.org Tue Nov 25 19:41:20 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 154F1D1B435 for ; Tue, 25 Nov 2003 23:41:18 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 55137-10 for ; Tue, 25 Nov 2003 19:40:48 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id E055DD1B432 for ; Tue, 25 Nov 2003 19:40:46 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAPNel19009471; Tue, 25 Nov 2003 18:40:47 -0500 (EST) To: josh@agliodbs.com Cc: pgsql-performance@postgresql.org Subject: Re: Wierd context-switching issue on Xeon In-reply-to: <200311251419.36771.josh@agliodbs.com> References: <200311251419.36771.josh@agliodbs.com> Comments: In-reply-to Josh Berkus message dated "Tue, 25 Nov 2003 14:19:36 -0800" Date: Tue, 25 Nov 2003 18:40:47 -0500 Message-ID: <9470.1069803647@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/259 X-Sequence-Number: 4819 Josh Berkus writes: > We're seeing some odd issues with hyperthreading-capable Xeons, whether or not > hyperthreading is enabled. Basically, when a small number of really-heavy > duty queries hit the system and push all of the CPUs to more than 70% used > (about 1/2 user & 1/2 kernel), the system goes to 100,000+ context switcthes > per second and performance degrades. Strictly a WAG ... but what this sounds like to me is disastrously bad behavior of the spinlock code under heavy contention. We thought we'd fixed the spinlock code for SMP machines awhile ago, but maybe hyperthreading opens some new vistas for misbehavior ... > I know that there's other Xeon users on this list ... has anyone else seen > anything like that? The machines are Dells running Red Hat 7.3. What Postgres version? Is it easy for you to try 7.4? If we were really lucky, the random-backoff algorithm added late in 7.4 development would cure this. If you can't try 7.4, or want to gather more data first, it would be good to try to confirm or disprove the theory that the context switches are coming from spinlock delays. If they are, they'd be coming from the select() calls in s_lock() in s_lock.c. Can you strace or something to see what kernel calls the context switches occur on? Another line of thought is that RH 7.3 is a long ways back, and it wasn't so very long ago that Linux still had lots of SMP bugs. Maybe what you really need is a kernel update? regards, tom lane From pgsql-performance-owner@postgresql.org Tue Nov 25 19:46:00 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 38D4AD1B44B for ; Tue, 25 Nov 2003 23:45:57 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 55083-08 for ; Tue, 25 Nov 2003 19:45:26 -0400 (AST) Received: from mta11.adelphia.net (mta11.adelphia.net [68.168.78.205]) by svr1.postgresql.org (Postfix) with ESMTP id B6AC4D1B44A for ; Tue, 25 Nov 2003 19:45:06 -0400 (AST) Received: from potentialtech.com ([68.68.113.33]) by mta11.adelphia.net (InterMail vM.5.01.06.05 201-253-122-130-105-20030824) with ESMTP id <20031125234513.JPLA1464.mta11.adelphia.net@potentialtech.com>; Tue, 25 Nov 2003 18:45:13 -0500 Message-ID: <3FC3E984.3020508@potentialtech.com> Date: Tue, 25 Nov 2003 18:45:08 -0500 From: Bill Moran User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.4) Gecko/20031005 X-Accept-Language: en-us, en MIME-Version: 1.0 To: schampailler@easynet.be Cc: pgsql-performance@postgresql.org Subject: Re: Impossibly slow DELETEs References: <200311252256.56433.schampailler@easynet.be> In-Reply-To: <200311252256.56433.schampailler@easynet.be> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/260 X-Sequence-Number: 4820 Stefan Champailler wrote: > Dear You all, > > (please tell me if this has already been discussed, I was unable to find any > convincing information) > > I'm developing a small application, tied to a PG 7.4 beta 5 (i didn't > upgrade). The DB i use is roughly 20 tales each of them containing at most 30 > records (I'm still in development). I can provide a whole dump if necessary. > I access the DB throug IODBC (Suse Linux 8.1), through PHP. The machine > everything runs on is 512M of Ram, 2.5GHz speed. So I assume it should be > blazingly fast. > > So here's my trouble : some DELETE statement take up to 1 minute to complete > (but not always, sometimes it's fast, sometimes it's that slow). Here's a > typical one : DELETE FROM response_bool WHERE response_id = '125' > The response_bool table has no foreing key and no index on response_id column. > No foreign key reference the response_bool table. There are 6 rows in the > table (given that size, I assumed that an index was not necessary). > > So 1 minute to complete look like I did something REALLY bad. > > It is my feeling that doing the same query with psql works without problem, > but I can't be sure. I think that last sentence is the crux of the problem. If you can establish for sure that the unreasonable delay is _only_ there when the command is issued through IODBC, then it's not a Postgresql problem. Out of curiosity, why are you using ODBC for PHP anyway? PHP has Postgresql libraries that work very well. I use them quite often without problems. -- Bill Moran Potential Technologies http://www.potentialtech.com From pgsql-performance-owner@postgresql.org Tue Nov 25 19:56:05 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id AC9A2D1B441 for ; Tue, 25 Nov 2003 23:56:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 59279-03 for ; Tue, 25 Nov 2003 19:55:33 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id E0BDDD1B446 for ; Tue, 25 Nov 2003 19:55:31 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id B12AE2092; Tue, 25 Nov 2003 18:55:33 -0500 (EST) To: schampailler@easynet.be Cc: pgsql-performance@postgresql.org Subject: Re: Impossibly slow DELETEs From: Neil Conway In-Reply-To: <200311252256.56433.schampailler@easynet.be> (Stefan Champailler's message of "Tue, 25 Nov 2003 22:56:56 +0100") References: <200311252256.56433.schampailler@easynet.be> Date: Tue, 25 Nov 2003 18:55:33 -0500 Message-ID: <87znekuhoq.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/262 X-Sequence-Number: 4822 Stefan Champailler writes: > So here's my trouble : some DELETE statement take up to 1 minute to > complete (but not always, sometimes it's fast, sometimes it's that > slow). Here's a typical one : DELETE FROM response_bool WHERE > response_id = '125' The response_bool table has no foreing key and > no index on response_id column. No foreign key reference the > response_bool table. I'm skeptical that PostgreSQL is causing the performance problem here -- 1 minute for a DELETE on a single-page table is absurdly slow. If you enable the log_min_duration_statement configuration variable, you should be able to get an idea of how long it actually takes PostgreSQL to execute each query -- do you see some 60 second queries in the log? What is the system load like when the query takes a long time? For example, `vmstat 1` output around this point in time would be helpful. Does PostgreSQL consume a lot of CPU time or do a lot of disk I/O? Can you confirm this problem using psql? > There are 6 rows in the table (given that size, I assumed that an > index was not necessary). That's a reasonable assumption. -Neil From pgsql-performance-owner@postgresql.org Tue Nov 25 20:37:58 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id E488AD1B49D for ; Wed, 26 Nov 2003 00:37:54 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 62520-08 for ; Tue, 25 Nov 2003 20:37:24 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id DBC64D1B44C for ; Tue, 25 Nov 2003 20:37:22 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAQ0bI19009858; Tue, 25 Nov 2003 19:37:18 -0500 (EST) To: Neil Conway Cc: schampailler@easynet.be, pgsql-performance@postgresql.org Subject: Re: Impossibly slow DELETEs In-reply-to: <87znekuhoq.fsf@mailbox.samurai.com> References: <200311252256.56433.schampailler@easynet.be> <87znekuhoq.fsf@mailbox.samurai.com> Comments: In-reply-to Neil Conway message dated "Tue, 25 Nov 2003 18:55:33 -0500" Date: Tue, 25 Nov 2003 19:37:18 -0500 Message-ID: <9857.1069807038@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/263 X-Sequence-Number: 4823 Neil Conway writes: >> There are 6 rows in the table (given that size, I assumed that an >> index was not necessary). > That's a reasonable assumption. But if he's updated those rows a few hundred thousand times and never VACUUMed, he could be having some problems ... regards, tom lane From pgsql-hackers-owner@postgresql.org Wed Nov 26 00:28:24 2003 X-Original-To: pgsql-hackers-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3D114D1B44E for ; Wed, 26 Nov 2003 04:28:19 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 84559-05 for ; Wed, 26 Nov 2003 00:27:47 -0400 (AST) Received: from mail.libertyrms.com (unknown [209.167.124.227]) by svr1.postgresql.org (Postfix) with ESMTP id 55315D1B492 for ; Wed, 26 Nov 2003 00:27:47 -0400 (AST) Received: from [10.1.2.130] (helo=dba2) by mail.libertyrms.com with esmtp (Exim 3.22 #3 (Debian)) id 1AOrHH-0007rD-00 for ; Tue, 25 Nov 2003 23:27:47 -0500 Received: by dba2 (Postfix, from userid 1019) id 045D7CC68; Tue, 25 Nov 2003 23:27:46 -0500 (EST) Date: Tue, 25 Nov 2003 23:27:46 -0500 From: Andrew Sullivan To: pgsql-hackers@postgresql.org Subject: Re: [PERFORM] More detail on settings for pgavd? Message-ID: <20031126042746.GB23015@libertyrms.info> Mail-Followup-To: Andrew Sullivan , pgsql-hackers@postgresql.org References: <200311181558.45847.josh@agliodbs.com> <200311210909.00978.josh@agliodbs.com> <3FBE8289.1040009@zeut.net> <200311211323.19003.josh@agliodbs.com> <15608.1069455896@sss.pgh.pa.us> <877k1t2pkq.fsf@stark.dyndns.tv> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <877k1t2pkq.fsf@stark.dyndns.tv> User-Agent: Mutt/1.5.4i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/1359 X-Sequence-Number: 47647 On Fri, Nov 21, 2003 at 07:51:17PM -0500, Greg Stark wrote: > The second vacuum waits for the lock to become available. If the > situation got really bad there could end up being a growing queue > of vacuums waiting. Those of us who have run into this know that "the situation got really bad" is earlier than one might think. And it can indeed cause some pretty pathological behaviour. A -- ---- Andrew Sullivan 204-4141 Yonge Street Afilias Canada Toronto, Ontario Canada M2P 2A8 +1 416 646 3304 x110 From pgsql-performance-owner@postgresql.org Wed Nov 26 02:07:14 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 2EE43D1B48D for ; Wed, 26 Nov 2003 06:07:11 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 92857-09 for ; Wed, 26 Nov 2003 02:06:40 -0400 (AST) Received: from smtp.pspl.co.in (www.pspl.co.in [202.54.11.65]) by svr1.postgresql.org (Postfix) with ESMTP id A9920D1B434 for ; Wed, 26 Nov 2003 02:06:37 -0400 (AST) Received: (from root@localhost) by smtp.pspl.co.in (8.12.9/8.12.9) id hAQ66gCZ030653 for ; Wed, 26 Nov 2003 11:36:42 +0530 Received: from persistent.co.in (daithan.intranet.pspl.co.in [192.168.7.161]) (authenticated bits=0) by persistent.co.in (8.12.9/8.12.9) with ESMTP id hAQ66eLp030627; Wed, 26 Nov 2003 11:36:40 +0530 Message-ID: <3FC442EA.7060300@persistent.co.in> Date: Wed, 26 Nov 2003 11:36:34 +0530 From: Shridhar Daithankar Organization: Persistent Systems Pvt. Ltd. User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Torsten Schulz Cc: pgsql-performance@postgresql.org Subject: Re: Optimize References: <3FC3D296.9040404@be-a-part.de> In-Reply-To: <3FC3D296.9040404@be-a-part.de> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/264 X-Sequence-Number: 4824 Torsten Schulz wrote: > Chester Kustarz wrote: >> On Mon, 24 Nov 2003, Torsten Schulz wrote: >>> shared_buffers = 5000 # 2*max_connections, min 16 >> that looks pretty small. that would only be 40MBytes (8k/page * >> 5000pages). >> http://www.varlena.com/GeneralBits/Tidbits/perf.html > Ok, thats it. I've set it to 51200, now it seems to be very fast. Whoa..That is too much. You acn get still better performance at something low like 10,000 or even 5000. Bumping up shared buffers stops being useful after a point and later it actually degrades the performance.. Shridhar From pgsql-performance-owner@postgresql.org Wed Nov 26 12:39:12 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 04EC6D1B45D for ; Wed, 26 Nov 2003 16:38:56 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 73348-01 for ; Wed, 26 Nov 2003 12:38:30 -0400 (AST) Received: from pcs-1.paccomsys.com (208.225.nwc.net [207.151.225.208]) by svr1.postgresql.org (Postfix) with ESMTP id 6100BD1B430 for ; Wed, 26 Nov 2003 12:38:23 -0400 (AST) Received: from paccomsys.com (mail.musicreports.com [64.161.179.34]) by pcs-1.paccomsys.com (8.12.5/8.12.5) with ESMTP id hAQGEi74011497 for ; Wed, 26 Nov 2003 08:14:45 -0800 Message-ID: <3FC4D703.5080909@paccomsys.com> Date: Wed, 26 Nov 2003 08:38:27 -0800 From: Roger Ging User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3) Gecko/20030708 Debian/1.3-4.lindows43 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: expression (functional) index use in joins Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/265 X-Sequence-Number: 4825 I just installed v7.4 and restored a database from v7.3.4. I have an index based on a function that the planner is using on the old version, but doing seq scans on left joins in the new version. I have run analyze on the table post restore. the query returns in less than 1 second on version 7.3.4 and takes over 10 seconds on version 7.4. Any help will be appreciated. Roger Ging Query: SELECT L.row_id FROM music.logfile L LEFT JOIN music.program P ON music.fn_mri_id_no_program(P.mri_id_no) = L.program_id WHERE L.station = UPPER('kabc')::VARCHAR AND L.air_date = '04/12/2002'::TIMESTAMP AND P.cutoff_date IS NULL ORDER BY L.chron_start,L.chron_end; planner results on 7.4: Sort (cost=17595.99..17608.23 rows=4894 width=12) Sort Key: l.chron_start, l.chron_end -> Merge Left Join (cost=17135.92..17296.07 rows=4894 width=12) Merge Cond: ("outer"."?column5?" = "inner"."?column3?") Filter: ("inner".cutoff_date IS NULL) -> Sort (cost=1681.69..1682.73 rows=414 width=21) Sort Key: (l.program_id)::text -> Index Scan using idx_logfile_station_air_date on logfile l (cost=0.00..1663.70 rows=414 width=21) Index Cond: (((station)::text = 'KABC'::text) AND (air_date = '2002-04-12 00:00:00'::timestamp without time zone)) -> Sort (cost=15454.22..15465.06 rows=4335 width=20) Sort Key: (music.fn_mri_id_no_program(p.mri_id_no))::text -> Seq Scan on program p (cost=0.00..15192.35 rows=4335 width=20) planner results on 7.3.4: Sort (cost=55765.51..55768.33 rows=1127 width=41) Sort Key: l.chron_start, l.chron_end -> Nested Loop (cost=0.00..55708.36 rows=1127 width=41) Filter: ("inner".cutoff_date IS NULL) -> Index Scan using idx_logfile_station_air_date on logfile l (cost=0.00..71.34 rows=17 width=21) Index Cond: ((station = 'KABC'::character varying) AND (air_date = '2002-04-12 00:00:00'::timestamp without time zone)) -> Index Scan using idx_program_mri_id_no_program on program p (cost=0.00..3209.16 rows=870 width=20) Index Cond: (music.fn_mri_id_no_program(p.mri_id_no) = "outer".program_id) table "Program" details: Column | Type | Modifiers ----------------+-----------------------------+----------- record_id | integer | title | character varying(40) | mri_id_no | character varying(8) | ascap_cat | character varying(1) | ascap_mult | numeric(5,3) | ascap_prod | character varying(10) | npa_ind | character varying(3) | non_inc_in | character varying(1) | as_pr_su | character varying(1) | as_1st_run | character varying(1) | as_cue_st | character varying(1) | bmi_cat | character varying(2) | bmi_mult | numeric(6,2) | bmi_prod | character varying(7) | year | integer | prog_type | character varying(1) | total_ep | integer | last_epis | character varying(3) | syndicator | character varying(6) | station | character varying(4) | syn_loc | character varying(1) | spdb_ver | character varying(4) | as_filed | character varying(4) | bmidb_ver | character varying(4) | cutoff_date | timestamp without time zone | effective_date | timestamp without time zone | program_id | character varying(5) | Indexes: "idx_program_mri_id_no" btree (mri_id_no) "idx_program_mri_id_no_program" btree (music.fn_mri_id_no_program(mri_id_no)) "idx_program_program_id" btree (program_id) "program_mri_id_no" btree (mri_id_no) "program_oid" btree (oid) From pgsql-performance-owner@postgresql.org Wed Nov 26 12:52:13 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A1225D1B430 for ; Wed, 26 Nov 2003 16:52:02 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 70459-08 for ; Wed, 26 Nov 2003 12:51:36 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id EF51BD1B44C for ; Wed, 26 Nov 2003 12:51:30 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hAQGpZNu086709 for ; Wed, 26 Nov 2003 16:51:35 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hAQGivrO085544 for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 16:44:57 GMT From: William Yu X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Maximum Possible Insert Performance? Date: Wed, 26 Nov 2003 08:44:59 -0800 Organization: Hub.Org Networking Services Lines: 42 Message-ID: References: <3FC19C2C.4050704@myrealbox.com> <200311240925.58070.josh@agliodbs.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Complaints-To: usenet@news.hub.org User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en-us, en In-Reply-To: <200311240925.58070.josh@agliodbs.com> To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/266 X-Sequence-Number: 4826 Josh Berkus wrote: > William, > > >>When my current job batch is done, I'll save a copy of the dir and give >>the WAL on ramdrive a test. And perhaps even buy a Sandisk at the local >>store and run that through the hooper. > > > We'll be interested in the results. The Sandisk won't be much of a > performance test; last I checked, their access speed was about 1/2 that of a > fast SCSI drive. But it could be a feasability test for the more expensive > RAMdrive approach. Some initial numbers. I simulated a CPU increase by underclocking the processors. Most of the time, performance does not scale linearly with clock speed but since I also underclocked the FSB and memory bandwidth with the CPU, it's nearly an exact match. 1.15GHz 6.14 1.53GHz 6.97 +33% CPU = +13.5% performance I then simulated adding a heapload of extra memory by running my job a second time. Unfortunately, to keep my 25GB DB mostly cached in memory, the word heapload is too accurate. Run 1 6.97 Run 2 7.99 +14% I popped in an extra IDE hard drive to store the WAL files and that boosted the numbers by a little. From looking at iostat, the ratio looked like 300K/s WAL for 1MB/s data. WAL+Data on same disk 6.97 WAL+Data separated 7.26 +4% I then tried to put the WAL directory onto a ramdisk. I turned off swapping, created a tmpfs mount point and copied the pg_xlog directory over. Everything looked fine as far as I could tell but Postgres just panic'd with a "file permissions" error. Anybody have thoughts to why tmpfs would not work? From pgsql-performance-owner@postgresql.org Wed Nov 26 13:12:03 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 214F5D1B45C for ; Wed, 26 Nov 2003 17:12:00 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 77504-01 for ; Wed, 26 Nov 2003 13:11:35 -0400 (AST) Received: from smtp.istop.com (dci.doncaster.on.ca [66.11.168.194]) by svr1.postgresql.org (Postfix) with ESMTP id 2E8A3D1B44D for ; Wed, 26 Nov 2003 13:11:29 -0400 (AST) Received: from stark.dyndns.tv (gsstark.mtl.istop.com [66.11.160.162]) by smtp.istop.com (Postfix) with ESMTP id 8D50636C4E; Wed, 26 Nov 2003 12:11:34 -0500 (EST) Received: from localhost ([127.0.0.1] helo=stark.dyndns.tv ident=foobar) by stark.dyndns.tv with smtp (Exim 3.36 #1 (Debian)) id 1AP3CQ-0000E4-00; Wed, 26 Nov 2003 12:11:34 -0500 To: Neil Conway Cc: schampailler@easynet.be, pgsql-performance@postgresql.org Subject: Re: Impossibly slow DELETEs References: <200311252256.56433.schampailler@easynet.be> <87znekuhoq.fsf@mailbox.samurai.com> In-Reply-To: <87znekuhoq.fsf@mailbox.samurai.com> From: Greg Stark Organization: The Emacs Conspiracy; member since 1992 Date: 26 Nov 2003 12:11:34 -0500 Message-ID: <87vfp7vyux.fsf@stark.dyndns.tv> Lines: 6 User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/267 X-Sequence-Number: 4827 Is it possible another connection has updated the record and not committed, and it takes a minute for the connection to time out and commit or roll back? -- greg From pgsql-performance-owner@postgresql.org Wed Nov 26 13:33:52 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id ED9C6D1B433 for ; Wed, 26 Nov 2003 17:33:50 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78416-08 for ; Wed, 26 Nov 2003 13:33:25 -0400 (AST) Received: from fuji.krosing.net (silmet.estpak.ee [194.126.97.78]) by svr1.postgresql.org (Postfix) with ESMTP id 7EE77D1B43A for ; Wed, 26 Nov 2003 13:33:18 -0400 (AST) Received: from fuji.krosing.net (localhost.localdomain [127.0.0.1]) by fuji.krosing.net (8.12.8/8.12.8) with ESMTP id hAQHXEhr003791; Wed, 26 Nov 2003 19:33:15 +0200 Received: (from hannu@localhost) by fuji.krosing.net (8.12.8/8.12.8/Submit) id hAQHX8rm003789; Wed, 26 Nov 2003 19:33:08 +0200 X-Authentication-Warning: fuji.krosing.net: hannu set sender to hannu@tm.ee using -f Subject: Re: why index scan not working when using 'like'? From: Hannu Krosing To: Tom Lane Cc: josh@agliodbs.com, LIANHE SHAO , pgsql-performance@postgresql.org In-Reply-To: <8754.1069795744@sss.pgh.pa.us> References: <3d83693d9fe0.3d9fe03d8369@jhmimail.jhmi.edu> <200311251151.40781.josh@agliodbs.com> <8754.1069795744@sss.pgh.pa.us> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1069867988.2749.26.camel@fuji.krosing.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Wed, 26 Nov 2003 19:33:08 +0200 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/268 X-Sequence-Number: 4828 Tom Lane kirjutas T, 25.11.2003 kell 23:29: > Josh Berkus writes: > > In regular text fields containing words, your problem is solvable with full > > text indexing (FTI). Unfortunately, FTI is not designed for arbitrary > > non-language strings. It could be adapted, but would require a lot of > > hacking. > > I'm not sure why you say that FTI isn't a usable solution. As long as > the gene symbols are separated by whitespace or some other non-letters > (eg, "foo mif bar" not "foomifbar"), I'd think FTI would work. If he wants to search on arbitrary substring, he could change tokeniser in FTI to produce trigrams, so that "foomifbar" would be indexed as if it were text "foo oom omi mif ifb fba bar" and search for things like %mifb% should first do a FTI search for "mif" AND "ifb" and then simple LIKE %mifb% to weed out something like "mififb". There are ways to use trigrams for 1 and 2 letter matches as well. ------------- Hannu From pgsql-performance-owner@postgresql.org Wed Nov 26 14:00:31 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1932BD1B42F for ; Wed, 26 Nov 2003 18:00:30 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 84719-04 for ; Wed, 26 Nov 2003 14:00:05 -0400 (AST) Received: from anchor-post-34.mail.demon.net (anchor-post-34.mail.demon.net [194.217.242.92]) by svr1.postgresql.org (Postfix) with ESMTP id 48422D1B433 for ; Wed, 26 Nov 2003 13:59:59 -0400 (AST) Received: from mwynhau.demon.co.uk ([193.237.186.96] helo=mainbox.archonet.com) by anchor-post-34.mail.demon.net with esmtp (Exim 3.35 #1) id 1AP3xL-000407-0Y; Wed, 26 Nov 2003 18:00:04 +0000 Received: by mainbox.archonet.com (Postfix, from userid 529) id AC2D116291; Wed, 26 Nov 2003 17:52:32 +0000 (GMT) Received: from client17.archonet.com (client17.archonet.com [192.168.1.17]) by mainbox.archonet.com (Postfix) with ESMTP id 56AE515D31; Wed, 26 Nov 2003 17:52:31 +0000 (GMT) From: Richard Huxton To: Roger Ging , pgsql-performance@postgresql.org Subject: Re: expression (functional) index use in joins Date: Wed, 26 Nov 2003 17:52:30 +0000 User-Agent: KMail/1.5 References: <3FC4D703.5080909@paccomsys.com> In-Reply-To: <3FC4D703.5080909@paccomsys.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311261752.30729.dev@archonet.com> X-Bogosity: No, tests=bogofilter, spamicity=0.000000, version=0.15.3 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/270 X-Sequence-Number: 4830 On Wednesday 26 November 2003 16:38, Roger Ging wrote: > I just installed v7.4 and restored a database from v7.3.4. [snip] Hmm - you seem to be getting different row estimates in the plan. Can you re-analyse both versions and post EXPLAIN ANALYSE rather than just EXPLAIN? > -> Seq Scan on program p (cost=0.00..15192.35 > rows=4335 width=20) > > planner results on 7.3.4: > > -> Index Scan using idx_program_mri_id_no_program on program > p (cost=0.00..3209.16 rows=870 width=20) -- Richard Huxton Archonet Ltd From pgsql-performance-owner@postgresql.org Wed Nov 26 13:56:24 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EE272D1B431 for ; Wed, 26 Nov 2003 17:56:22 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 85112-01 for ; Wed, 26 Nov 2003 13:55:58 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id D975ED1B489 for ; Wed, 26 Nov 2003 13:55:51 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAQHtu19016110; Wed, 26 Nov 2003 12:55:57 -0500 (EST) To: William Yu Cc: pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? In-reply-to: References: <3FC19C2C.4050704@myrealbox.com> <200311240925.58070.josh@agliodbs.com> Comments: In-reply-to William Yu message dated "Wed, 26 Nov 2003 08:44:59 -0800" Date: Wed, 26 Nov 2003 12:55:56 -0500 Message-ID: <16109.1069869356@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/269 X-Sequence-Number: 4829 William Yu writes: > I then tried to put the WAL directory onto a ramdisk. I turned off > swapping, created a tmpfs mount point and copied the pg_xlog directory > over. Everything looked fine as far as I could tell but Postgres just > panic'd with a "file permissions" error. Anybody have thoughts to why > tmpfs would not work? I'd say you got the file or directory ownership or permissions wrong. regards, tom lane From pgsql-performance-owner@postgresql.org Wed Nov 26 14:22:11 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 0754ED1B439 for ; Wed, 26 Nov 2003 18:22:03 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 86685-05 for ; Wed, 26 Nov 2003 14:21:37 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by svr1.postgresql.org (Postfix) with ESMTP id 5F0A8D1B432 for ; Wed, 26 Nov 2003 14:21:31 -0400 (AST) Received: from news.hub.org (news.hub.org [200.46.204.72]) by news.hub.org (8.12.9/8.12.9) with ESMTP id hAQILaNw003602 for ; Wed, 26 Nov 2003 18:21:36 GMT (envelope-from news@news.hub.org) Received: (from news@localhost) by news.hub.org (8.12.9/8.12.9/Submit) id hAQI3iuq000563 for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 18:03:45 GMT From: William Yu X-Newsgroups: comp.databases.postgresql.performance Subject: Re: Maximum Possible Insert Performance? Date: Wed, 26 Nov 2003 10:03:47 -0800 Organization: Hub.Org Networking Services Lines: 14 Message-ID: References: <3FC19C2C.4050704@myrealbox.com> <200311240925.58070.josh@agliodbs.com> <16109.1069869356@sss.pgh.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Complaints-To: usenet@news.hub.org User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en-us, en In-Reply-To: <16109.1069869356@sss.pgh.pa.us> To: pgsql-performance@postgresql.org X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/271 X-Sequence-Number: 4831 Tom Lane wrote: > William Yu writes: > >>I then tried to put the WAL directory onto a ramdisk. I turned off >>swapping, created a tmpfs mount point and copied the pg_xlog directory >>over. Everything looked fine as far as I could tell but Postgres just >>panic'd with a "file permissions" error. Anybody have thoughts to why >>tmpfs would not work? > > > I'd say you got the file or directory ownership or permissions wrong. I did a mv instead of a cp which duplicates ownership & permissions exactly. From pgsql-performance-owner@postgresql.org Wed Nov 26 14:39:57 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 360F3D1B43F for ; Wed, 26 Nov 2003 18:39:56 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 86685-10 for ; Wed, 26 Nov 2003 14:39:31 -0400 (AST) Received: from pcs-1.paccomsys.com (208.225.nwc.net [207.151.225.208]) by svr1.postgresql.org (Postfix) with ESMTP id C87F5D1B431 for ; Wed, 26 Nov 2003 14:39:23 -0400 (AST) Received: from paccomsys.com (mail.musicreports.com [64.161.179.34]) by pcs-1.paccomsys.com (8.12.5/8.12.5) with ESMTP id hAQIFm74012678 for ; Wed, 26 Nov 2003 10:15:49 -0800 Message-ID: <3FC4F360.2090609@paccomsys.com> Date: Wed, 26 Nov 2003 10:39:28 -0800 From: Roger Ging User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3) Gecko/20030708 Debian/1.3-4.lindows43 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Followup - expression (functional) index use in joins Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/272 X-Sequence-Number: 4832 version 7.4 results: explain analyse SELECT L.row_id FROM music.logfile L LEFT JOIN music.program P ON music.fn_mri_id_no_program(P.mri_id_no) = L.program_id WHERE L.station = UPPER('kabc')::VARCHAR AND L.air_date = '04/12/2002'::TIMESTAMP AND P.cutoff_date IS NULL ORDER BY L.chron_start,L.chron_end; QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------------------------------------------- Sort (cost=17595.99..17608.23 rows=4894 width=12) (actual time=8083.719..8083.738 rows=30 loops=1) Sort Key: l.chron_start, l.chron_end -> Merge Left Join (cost=17135.92..17296.07 rows=4894 width=12) (actual time=7727.590..8083.349 rows=30 loops=1) Merge Cond: ("outer"."?column5?" = "inner"."?column3?") Filter: ("inner".cutoff_date IS NULL) -> Sort (cost=1681.69..1682.73 rows=414 width=21) (actual time=1.414..1.437 rows=30 loops=1) Sort Key: (l.program_id)::text -> Index Scan using idx_logfile_station_air_date on logfile l (cost=0.00..1663.70 rows=414 width=21) (actual time=0.509..1.228 rows=30 loops=1) Index Cond: (((station)::text = 'KABC'::text) AND (air_date = '2002-04-12 00:00:00'::timestamp without time zone)) -> Sort (cost=15454.22..15465.06 rows=4335 width=20) (actual time=7718.612..7869.874 rows=152779 loops=1) Sort Key: (music.fn_mri_id_no_program(p.mri_id_no))::text -> Seq Scan on program p (cost=0.00..15192.35 rows=4335 width=20) (actual time=109.045..1955.882 rows=173998 loops=1) Total runtime: 8194.290 ms (13 rows) version 7.3 results: explain analyse SELECT L.row_id FROM music.logfile L LEFT JOIN music.program P ON music.fn_mri_id_no_program(P.mri_id_no) = L.program_id WHERE L.station = UPPER('kabc')::VARCHAR AND L.air_date = '04/12/2002'::TIMESTAMP AND P.cutoff_date IS NULL ORDER BY L.chron_start,L.chron_end; QUERY PLAN ---------------------------------------------------------------------------------------------------------------------------------------------------------- Sort (cost=55765.51..55768.33 rows=1127 width=41) (actual time=7.74..7.75 rows=30 loops=1) Sort Key: l.chron_start, l.chron_end -> Nested Loop (cost=0.00..55708.36 rows=1127 width=41) (actual time=0.21..7.62 rows=30 loops=1) Filter: ("inner".cutoff_date IS NULL) -> Index Scan using idx_logfile_station_air_date on logfile l (cost=0.00..71.34 rows=17 width=21) (actual time=0.14..0.74 rows=30 loops=1) Index Cond: ((station = 'KABC'::character varying) AND (air_date = '2002-04-12 00:00:00'::timestamp without time zone)) -> Index Scan using idx_program_mri_id_no_program on program p (cost=0.00..3209.16 rows=870 width=20) (actual time=0.05..0.22 rows=9 loops=30) Index Cond: (music.fn_mri_id_no_program(p.mri_id_no) = "outer".program_id) Total runtime: 7.86 msec From pgsql-performance-owner@postgresql.org Wed Nov 26 14:54:54 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1F540D1B430 for ; Wed, 26 Nov 2003 18:54:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 89528-08 for ; Wed, 26 Nov 2003 14:54:28 -0400 (AST) Received: from rlx13.zapatec.com (66-117-144-213.zapatec.lmi.net [66.117.144.213]) by svr1.postgresql.org (Postfix) with ESMTP id 58FACD1B437 for ; Wed, 26 Nov 2003 14:54:21 -0400 (AST) Received: from rlx11.zapatec.com (rlx11.pr.zapatec.com [192.168.1.132]) by rlx13.zapatec.com (Postfix) with ESMTP id AF44CA941 for ; Wed, 26 Nov 2003 10:54:22 -0800 (PST) Received: (from dror@localhost) by rlx11.zapatec.com (8.12.3/8.12.3/Submit) id hAQIsL6p043232 for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 10:54:21 -0800 (PST) (envelope-from dror) Date: Wed, 26 Nov 2003 10:54:21 -0800 From: Dror Matalon To: pgsql-performance@postgresql.org Subject: Re: Maximum Possible Insert Performance? Message-ID: <20031126185421.GN30893@rlx11.zapatec.com> References: <3FC19C2C.4050704@myrealbox.com> <200311240925.58070.josh@agliodbs.com> <16109.1069869356@sss.pgh.pa.us> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/273 X-Sequence-Number: 4833 But the permissions of the base ramdisk might be wrong. I'd su to the user that you run postgres as (probably postgres), and make sure that you can go to the directory where the log and the database files are and make sure you can see the files. On Wed, Nov 26, 2003 at 10:03:47AM -0800, William Yu wrote: > Tom Lane wrote: > >William Yu writes: > > > >>I then tried to put the WAL directory onto a ramdisk. I turned off > >>swapping, created a tmpfs mount point and copied the pg_xlog directory > >>over. Everything looked fine as far as I could tell but Postgres just > >>panic'd with a "file permissions" error. Anybody have thoughts to why > >>tmpfs would not work? > > > > > >I'd say you got the file or directory ownership or permissions wrong. > > I did a mv instead of a cp which duplicates ownership & permissions exactly. > > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster -- Dror Matalon Zapatec Inc 1700 MLK Way Berkeley, CA 94709 http://www.fastbuzz.com http://www.zapatec.com From pgsql-performance-owner@postgresql.org Wed Nov 26 15:18:21 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 46D2FD1B445 for ; Wed, 26 Nov 2003 19:13:26 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 93073-07 for ; Wed, 26 Nov 2003 15:13:01 -0400 (AST) Received: from anchor-post-33.mail.demon.net (anchor-post-33.mail.demon.net [194.217.242.91]) by svr1.postgresql.org (Postfix) with ESMTP id B3BA9D1B48F for ; Wed, 26 Nov 2003 15:11:59 -0400 (AST) Received: from mwynhau.demon.co.uk ([193.237.186.96] helo=mainbox.archonet.com) by anchor-post-33.mail.demon.net with esmtp (Exim 3.35 #1) id 1AP553-000GEZ-0X; Wed, 26 Nov 2003 19:12:05 +0000 Received: by mainbox.archonet.com (Postfix, from userid 529) id 7E10B16570; Wed, 26 Nov 2003 19:12:04 +0000 (GMT) Received: from client17.archonet.com (client17.archonet.com [192.168.1.17]) by mainbox.archonet.com (Postfix) with ESMTP id A2D3D15D31; Wed, 26 Nov 2003 19:12:02 +0000 (GMT) From: Richard Huxton To: Roger Ging , pgsql-performance@postgresql.org Subject: Re: Followup - expression (functional) index use in joins Date: Wed, 26 Nov 2003 19:12:01 +0000 User-Agent: KMail/1.5 References: <3FC4F360.2090609@paccomsys.com> In-Reply-To: <3FC4F360.2090609@paccomsys.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311261912.01286.dev@archonet.com> X-Bogosity: No, tests=bogofilter, spamicity=0.000000, version=0.15.3 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/274 X-Sequence-Number: 4834 On Wednesday 26 November 2003 18:39, Roger Ging wrote: > version 7.4 results: > > explain analyse SELECT L.row_id FROM music.logfile L LEFT JOIN > music.program P ON > music.fn_mri_id_no_program(P.mri_id_no) = L.program_id > WHERE L.station = UPPER('kabc')::VARCHAR > AND L.air_date = '04/12/2002'::TIMESTAMP > AND P.cutoff_date IS NULL > ORDER BY L.chron_start,L.chron_end; > -> Seq Scan on program p (cost=0.00..15192.35 > rows=4335 width=20) (actual time=109.045..1955.882 rows=173998 loops=1) The estimated number of rows here (4335) is *way* off (173998 actually). If you only had 4335 rows, then this might be a more sensible plan. First step is to run: VACUUM ANALYSE program; Then, check the definition of your function fn_mri_id_no_program() and make sure it is marked immutable/stable (depending on what it does) and that it's returning a varchar. -- Richard Huxton Archonet Ltd From pgsql-performance-owner@postgresql.org Wed Nov 26 16:03:34 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 800BBD1B439 for ; Wed, 26 Nov 2003 20:03:33 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 03910-07 for ; Wed, 26 Nov 2003 16:03:09 -0400 (AST) Received: from jhuml1.jhmi.edu (jhuml1.jhmi.edu [162.129.234.20]) by svr1.postgresql.org (Postfix) with ESMTP id 1FAF7D1B446 for ; Wed, 26 Nov 2003 16:03:02 -0400 (AST) Received: from jhuml1.jhmi.edu (jhuml1.jhmi.edu [162.129.234.20]) by jhuml1.jhmi.edu (PMDF V6.2-X17 #30839) with SMTP id <0HOZ00IUP69ZN1@jhuml1.jhmi.edu> for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 15:02:53 -0500 (EST) Received: from jhuml1.jhmi.edu ([162.129.234.20]) by jhuml1.jhmi.edu (SAVSMTP 3.1.0.29) with SMTP id M2003112615025324392 for ; Wed, 26 Nov 2003 15:02:53 -0500 Received: from jhmimail.jhmi.edu (jhem2.jhmi.edu [162.129.8.23]) by jhuml1.jhmi.edu (PMDF V6.2-X17 #30839) with ESMTP id <0HOZ00IQ26CSN1@jhuml1.jhmi.edu> for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 15:02:52 -0500 (EST) Received: from [162.129.178.60] by jhmimail.jhmi.edu (mshttpd); Wed, 26 Nov 2003 20:06:02 +0000 (GMT) Date: Wed, 26 Nov 2003 20:06:02 +0000 (GMT) From: LIANHE SHAO Subject: For full text indexing, which is better, tsearch2 or fulltextindex To: pgsql-performance@postgresql.org Message-id: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> MIME-version: 1.0 X-Mailer: iPlanet Messenger Express 5.2 HotFix 1.17 (built Jun 23 2003) Content-type: text/plain; charset=us-ascii Content-language: en Content-transfer-encoding: 7bit Content-disposition: inline X-Accept-Language: en X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/275 X-Sequence-Number: 4835 Hi all, Which one is better (performance/easier to use), tsearch2 or fulltextindex? there is an example how to use fulltextindex in the techdocs, but I checked the contrib/fulltextindex package, there is a WARNING that fulltextindex is much slower than tsearch2. but tsearch2 seems complex to use, and I can not find a good example. Which one I should use? Any suggestions? thanks and Regards, William ----- Original Message ----- From: Hannu Krosing Date: Wednesday, November 26, 2003 5:33 pm Subject: Re: [PERFORM] why index scan not working when using 'like'? > Tom Lane kirjutas T, 25.11.2003 kell 23:29: > > Josh Berkus writes: > > > In regular text fields containing words, your problem is > solvable with full > > > text indexing (FTI). Unfortunately, FTI is not designed for > arbitrary > > > non-language strings. It could be adapted, but would require a > lot of > > > hacking. > > > > I'm not sure why you say that FTI isn't a usable solution. As > long as > > the gene symbols are separated by whitespace or some other non- > letters> (eg, "foo mif bar" not "foomifbar"), I'd think FTI would > work. > If he wants to search on arbitrary substring, he could change > tokeniserin FTI to produce trigrams, so that "foomifbar" would be > indexed as if > it were text "foo oom omi mif ifb fba bar" and search for things like > %mifb% should first do a FTI search for "mif" AND "ifb" and then > simpleLIKE %mifb% to weed out something like "mififb". > > There are ways to use trigrams for 1 and 2 letter matches as well. > > ------------- > Hannu > > > ---------------------------(end of broadcast)----------------------- > ---- > TIP 3: if posting/reading through Usenet, please send an appropriate > subscribe-nomail command to majordomo@postgresql.org so that > your message can get through to the mailing list cleanly > From pgsql-performance-owner@postgresql.org Wed Nov 26 16:14:53 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 611CAD1B445 for ; Wed, 26 Nov 2003 20:14:52 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 71751-05 for ; Wed, 26 Nov 2003 16:14:28 -0400 (AST) Received: from penguin.goodinassociates.com (unknown [63.150.225.202]) by svr1.postgresql.org (Postfix) with ESMTP id 76276D1B434 for ; Wed, 26 Nov 2003 16:14:20 -0400 (AST) Received: from [192.168.14.183] (bluejay.goodinassociates.com [192.168.14.183]) by penguin.goodinassociates.com (8.12.8/linuxconf) with ESMTP id hAQKEBna013924 for ; Wed, 26 Nov 2003 14:14:11 -0600 Subject: cross table indexes or something? From: Jeremiah Jahn To: postgres performance Content-Type: text/plain Message-Id: <1069877651.22346.13.camel@bluejay.goodinassociates.com> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 (1.4.5-7) Date: Wed, 26 Nov 2003 14:14:11 -0600 Content-Transfer-Encoding: 7bit X-RAVMilter-Version: 8.4.4(snapshot 20030410) (penguin.goodinassociates.com) X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/276 X-Sequence-Number: 4836 I was wondering if there is something I can do that would act similar to a index over more than one table. I have about 3 million people in my DB at the moment, they all have roles, and many of them have more than one name. for example, a Judge will only have one name, but a Litigant could have multiple aliases. Things go far to slow when I do a query on a judge named smith. Does any one know a possible way to speed this up? I would think that In a perfect world there would be a way to create an index on commonly used joins, or something of that nature. I've tried partial indexes, but the optimizer feels that it would be quicker to do an index scan for smith% then join using the pkey of the person to get their role. For litigants, this makes since, for non-litigants, this doesn't. thanx for any insight, -jj- the basic schema actor actor_id PK role_class_code identity actor_id FK identity_id PK full_name event event_date_time event_id PK event_actor event_id FK actor_id FK explain select distinct actor.actor_id,court.id,court.name,role_class_code,full_name from actor,identity,court,event,event_actor where role_class_code = 'Judge' and full_name like 'SMITH%' and identity.actor_id = actor.actor_id and identity.court_ori = actor.court_ori and actor.court_ori = court.id and actor.actor_id = event_actor.actor_id and event_actor.event_id = event.event_id and event_date_time > '20021126' order by full_name; QUERY PLAN ------------------------------------------------------------------------------------------------------------------------------------------------------------------ Unique (cost=726.57..726.58 rows=1 width=92) -> Sort (cost=726.57..726.57 rows=1 width=92) Sort Key: identity.full_name, actor.actor_id, court.id, court.name, actor.role_class_code -> Nested Loop (cost=3.02..726.56 rows=1 width=92) -> Nested Loop (cost=3.02..720.72 rows=1 width=144) -> Nested Loop (cost=3.02..9.62 rows=1 width=117) Join Filter: (("outer".court_ori)::text = ("inner".court_ori)::text) -> Hash Join (cost=3.02..4.18 rows=1 width=93) Hash Cond: (("outer".id)::text = ("inner".court_ori)::text) -> Seq Scan on court (cost=0.00..1.10 rows=10 width=34) -> Hash (cost=3.01..3.01 rows=1 width=59) -> Index Scan using name_speed on identity (cost=0.00..3.01 rows=1 width=59) Index Cond: (((full_name)::text >= 'SMITH'::character varying) AND ((full_name)::text < 'SMITI'::character varying)) Filter: ((full_name)::text ~~ 'SMITH%'::text) -> Index Scan using actor_speed on actor (cost=0.00..5.43 rows=1 width=50) Index Cond: (("outer".actor_id)::text = (actor.actor_id)::text) Filter: ((role_class_code)::text = 'Judge'::text) -> Index Scan using event_actor_speed on event_actor (cost=0.00..695.15 rows=1275 width=73) Index Cond: ((event_actor.actor_id)::text = ("outer".actor_id)::text) -> Index Scan using event_pkey on event (cost=0.00..5.83 rows=1 width=52) Index Cond: (("outer".event_id)::text = (event.event_id)::text) Filter: (event_date_time > '20021126'::bpchar) -- "You can't make a program without broken egos." -- Jeremiah Jahn From pgsql-performance-owner@postgresql.org Wed Nov 26 16:19:19 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C5F2ED1B439 for ; Wed, 26 Nov 2003 20:19:17 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 72256-04 for ; Wed, 26 Nov 2003 16:18:53 -0400 (AST) Received: from jhuml1.jhmi.edu (jhuml1.jhmi.edu [162.129.234.20]) by svr1.postgresql.org (Postfix) with ESMTP id 11879D1B43E for ; Wed, 26 Nov 2003 16:18:46 -0400 (AST) Received: from jhuml1.jhmi.edu (jhuml1.jhmi.edu [162.129.234.20]) by jhuml1.jhmi.edu (PMDF V6.2-X17 #30839) with SMTP id <0HOZ003VR6ZNE1@jhuml1.jhmi.edu> for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 15:18:52 -0500 (EST) Received: from jhuml1.jhmi.edu ([162.129.234.20]) by jhuml1.jhmi.edu (SAVSMTP 3.1.0.29) with SMTP id M2003112615185228023 for ; Wed, 26 Nov 2003 15:18:52 -0500 Received: from jhmimail.jhmi.edu (jhem2.jhmi.edu [162.129.8.23]) by jhuml1.jhmi.edu (PMDF V6.2-X17 #30839) with ESMTP id <0HOZ008FJ73GMX@jhuml1.jhmi.edu> for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 15:18:52 -0500 (EST) Received: from [162.129.178.60] by jhmimail.jhmi.edu (mshttpd); Wed, 26 Nov 2003 20:22:01 +0000 (GMT) Date: Wed, 26 Nov 2003 20:22:01 +0000 (GMT) From: LIANHE SHAO Subject: very large db performance question To: pgsql-performance@postgresql.org Message-id: <4cbacd4cb3c2.4cb3c24cbacd@jhmimail.jhmi.edu> MIME-version: 1.0 X-Mailer: iPlanet Messenger Express 5.2 HotFix 1.17 (built Jun 23 2003) Content-type: text/plain; charset=us-ascii Content-language: en Content-transfer-encoding: 7bit Content-disposition: inline X-Accept-Language: en X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/277 X-Sequence-Number: 4837 Hello All, We will have a very large database to store microarray data (may exceed 80-100G some day). now we have 1G RAM, 2G Hz Pentium 4, 1 CPU. and enough hard disk. I never touched such large database before. I ask several dbas if the hardware is ok, some said it is ok for the query, but I am not so convinced. Because I check the mailing list and learned that it is not unreasonable to take several minutes to do the query. But I want to query to be as fast as possible. Could anybody tell me that our hardware is an issue or not? do we really need better hardware to make real difference? Regards, William From pgsql-performance-owner@postgresql.org Wed Nov 26 17:29:47 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 388D0D1B43F for ; Wed, 26 Nov 2003 21:29:44 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78896-07 for ; Wed, 26 Nov 2003 17:29:19 -0400 (AST) Received: from pcs-1.paccomsys.com (208.225.nwc.net [207.151.225.208]) by svr1.postgresql.org (Postfix) with ESMTP id 6D2D8D1B482 for ; Wed, 26 Nov 2003 17:29:11 -0400 (AST) Received: from paccomsys.com (mail.musicreports.com [64.161.179.34]) by pcs-1.paccomsys.com (8.12.5/8.12.5) with ESMTP id hAQL5f74014340 for ; Wed, 26 Nov 2003 13:05:42 -0800 Message-ID: <3FC51B2C.2030700@paccomsys.com> Date: Wed, 26 Nov 2003 13:29:16 -0800 From: Roger Ging User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3) Gecko/20030708 Debian/1.3-4.lindows43 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: Re: Followup - expression (functional) index use in joins References: <3FC4F360.2090609@paccomsys.com> <200311261912.01286.dev@archonet.com> In-Reply-To: <200311261912.01286.dev@archonet.com> Content-Type: multipart/alternative; boundary="------------060905040807020201000309" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/278 X-Sequence-Number: 4838 This is a multi-part message in MIME format. --------------060905040807020201000309 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Ran vacuum analyse on both program and logfile tables. Estimates are more in line with reality now, but query still takes 10 seconds on v7.4 and 10 ms on v7.3. Function is marked as immutable and returns varchar(5). I am wondering why the planner would choose a merge join (v7.4) as opposed to a nested loop (v7.3) given the small number of rows in the top level table (logfile) based upon the where clause ( L.air_date = '04/12/2002'::TIMESTAMP ) there are typically only 30 rows per station/air_date. What am I missing here? Richard Huxton wrote: >On Wednesday 26 November 2003 18:39, Roger Ging wrote: > > >>version 7.4 results: >> >>explain analyse SELECT L.row_id FROM music.logfile L LEFT JOIN >>music.program P ON >>music.fn_mri_id_no_program(P.mri_id_no) = L.program_id >>WHERE L.station = UPPER('kabc')::VARCHAR >>AND L.air_date = '04/12/2002'::TIMESTAMP >>AND P.cutoff_date IS NULL >>ORDER BY L.chron_start,L.chron_end; >> >> > > > >> -> Seq Scan on program p (cost=0.00..15192.35 >>rows=4335 width=20) (actual time=109.045..1955.882 rows=173998 loops=1) >> >> > >The estimated number of rows here (4335) is *way* off (173998 actually). If >you only had 4335 rows, then this might be a more sensible plan. > >First step is to run: > VACUUM ANALYSE program; >Then, check the definition of your function fn_mri_id_no_program() and make >sure it is marked immutable/stable (depending on what it does) and that it's >returning a varchar. > > > > --------------060905040807020201000309 Content-Type: text/html; charset=us-ascii Content-Transfer-Encoding: 7bit Ran vacuum analyse on both program and logfile tables.  Estimates are more in line with reality now, but query still takes 10 seconds on v7.4 and 10 ms on v7.3.  Function is marked as immutable and returns varchar(5).  I am wondering why the planner would choose a merge join (v7.4) as opposed to a nested loop (v7.3) given the small number of rows in the top level table (logfile) based upon the where clause ( 
L.air_date = '04/12/2002'::TIMESTAMP
)
there are typically only 30 rows per station/air_date.  What am I missing here?

Richard Huxton wrote:
On Wednesday 26 November 2003 18:39, Roger Ging wrote:
  
version 7.4 results:

explain analyse SELECT L.row_id FROM music.logfile L LEFT JOIN
music.program P ON
music.fn_mri_id_no_program(P.mri_id_no) = L.program_id
WHERE  L.station = UPPER('kabc')::VARCHAR
AND L.air_date = '04/12/2002'::TIMESTAMP
AND P.cutoff_date IS NULL
ORDER BY L.chron_start,L.chron_end;
    

  
                ->  Seq Scan on program p  (cost=0.00..15192.35
rows=4335 width=20) (actual time=109.045..1955.882 rows=173998 loops=1)
    

The estimated number of rows here (4335) is *way* off (173998 actually). If 
you only had 4335 rows, then this might be a more sensible plan.

First step is to run:
  VACUUM ANALYSE program;
Then, check the definition of your function fn_mri_id_no_program() and make 
sure it is marked immutable/stable (depending on what it does) and that it's 
returning a varchar.
--------------060905040807020201000309-- From pgsql-performance-owner@postgresql.org Wed Nov 26 18:04:01 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 4A266D1B43C for ; Wed, 26 Nov 2003 22:03:56 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 86105-02 for ; Wed, 26 Nov 2003 18:03:32 -0400 (AST) Received: from bob.samurai.com (bob.samurai.com [205.207.28.75]) by svr1.postgresql.org (Postfix) with ESMTP id 749B2D1B48A for ; Wed, 26 Nov 2003 18:03:24 -0400 (AST) Received: from tokyo.samurai.com (d226-89-59.home.cgocable.net [24.226.89.59]) by bob.samurai.com (Postfix) with ESMTP id 19EDF1F40; Wed, 26 Nov 2003 17:03:31 -0500 (EST) To: LIANHE SHAO Cc: pgsql-performance@postgresql.org Subject: Re: very large db performance question From: Neil Conway In-Reply-To: <4cbacd4cb3c2.4cb3c24cbacd@jhmimail.jhmi.edu> (LIANHE SHAO's message of "Wed, 26 Nov 2003 20:22:01 +0000 (GMT)") References: <4cbacd4cb3c2.4cb3c24cbacd@jhmimail.jhmi.edu> Date: Wed, 26 Nov 2003 17:03:31 -0500 Message-ID: <87u14q6b4c.fsf@mailbox.samurai.com> User-Agent: Gnus/5.1002 (Gnus v5.10.2) XEmacs/21.4 (Reasonable Discussion, linux) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/279 X-Sequence-Number: 4839 LIANHE SHAO writes: > We will have a very large database to store microarray data (may > exceed 80-100G some day). now we have 1G RAM, 2G Hz Pentium 4, 1 > CPU. and enough hard disk. > Could anybody tell me that our hardware is an issue or not? IMHO the size of the DB is less relevant than the query workload. For example, if you're storying 100GB of data but only doing a single index scan on it every 10 seconds, any modern machine with enough HD space should be fine. If you give us an idea of the # of queries you expect per second, the approximate mix of reads and writes, and some idea of how complex the queries are, we might be able to give you some better advice. -Neil From pgsql-performance-owner@postgresql.org Wed Nov 26 18:24:25 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 3CFECD1B430 for ; Wed, 26 Nov 2003 22:24:14 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 85810-08 for ; Wed, 26 Nov 2003 18:23:50 -0400 (AST) Received: from appsrv1.nuvergence.com (appsrv1.mgamble.ca [64.56.239.120]) by svr1.postgresql.org (Postfix) with ESMTP id 5F71CD1B43F for ; Wed, 26 Nov 2003 18:23:38 -0400 (AST) Received: from localhost (localhost [127.0.0.1]) by appsrv1.nuvergence.com (Postfix) with ESMTP id 4E54330AE20 for ; Wed, 26 Nov 2003 17:23:14 -0500 (EST) Received: from host-130.in3healthsystems.com ( [host-130.in3healthsystems.com]) as user marc@redboxdata.com@appsrv1.simpledomains.net by webmail.nuvergence.com with HTTP; Wed, 26 Nov 2003 17:23:14 -0500 Message-ID: <1069885394.3fc527d23560e@webmail.nuvergence.com> Date: Wed, 26 Nov 2003 17:23:14 -0500 From: "Marc A. Leith" To: postgres performance Subject: Re: cross table indexes or something? References: <1069877651.22346.13.camel@bluejay.goodinassociates.com> In-Reply-To: <1069877651.22346.13.camel@bluejay.goodinassociates.com> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.0 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/280 X-Sequence-Number: 4840 Sybase IQ lets you build "joined indexsets". This is amazing but pricey and really intended more for Data Warehousing than OLTP, although they did release a version which permitted writes on-the-fly. (This was implemented using a multi-concurrency solution much like PostreSQL uses.) It essentially pre-joined the data. Marc A. Leith redboxdata inc. E-mail:mleith@redboxdata.com Quoting Jeremiah Jahn : > I was wondering if there is something I can do that would act similar to > a index over more than one table. > > I have about 3 million people in my DB at the moment, they all have > roles, and many of them have more than one name. > > for example, a Judge will only have one name, but a Litigant could have > multiple aliases. Things go far to slow when I do a query on a judge > named smith. Does any one know a possible way to speed this up? > > I would think that In a perfect world there would be a way to create an > index on commonly used joins, or something of that nature. I've tried > partial indexes, but the optimizer feels that it would be quicker to do > an index scan for smith% then join using the pkey of the person to get > their role. For litigants, this makes since, for non-litigants, this > doesn't. > > thanx for any insight, > -jj- > > the basic schema > > actor > actor_id PK > role_class_code > > identity > actor_id FK > identity_id PK > full_name > > event > event_date_time > event_id PK > > event_actor > event_id FK > actor_id FK > > > explain select distinct > actor.actor_id,court.id,court.name,role_class_code,full_name from > actor,identity,court,event,event_actor where role_class_code = 'Judge' and > full_name like 'SMITH%' and identity.actor_id = actor.actor_id and > identity.court_ori = actor.court_ori and actor.court_ori = court.id and > actor.actor_id = event_actor.actor_id and event_actor.event_id = > event.event_id and event_date_time > '20021126' order by full_name; > > QUERY PLAN > ------------------------------------------------------------------------------ -------------------------------------------------------------------------------- ---- > Unique (cost=726.57..726.58 rows=1 width=92) > -> Sort (cost=726.57..726.57 rows=1 width=92) > Sort Key: identity.full_name, actor.actor_id, court.id, court.name, > actor.role_class_code > -> Nested Loop (cost=3.02..726.56 rows=1 width=92) > -> Nested Loop (cost=3.02..720.72 rows=1 width=144) > -> Nested Loop (cost=3.02..9.62 rows=1 width=117) > Join Filter: (("outer".court_ori)::text = > ("inner".court_ori)::text) > -> Hash Join (cost=3.02..4.18 rows=1 width=93) > Hash Cond: (("outer".id)::text = > ("inner".court_ori)::text) > -> Seq Scan on court (cost=0.00..1.10 > rows=10 width=34) > -> Hash (cost=3.01..3.01 rows=1 width=59) > -> Index Scan using name_speed on > identity (cost=0.00..3.01 rows=1 width=59) > Index Cond: (((full_name)::text > >= 'SMITH'::character varying) AND ((full_name)::text < 'SMITI'::character > varying)) > Filter: ((full_name)::text ~~ > 'SMITH%'::text) > -> Index Scan using actor_speed on actor > (cost=0.00..5.43 rows=1 width=50) > Index Cond: (("outer".actor_id)::text = > (actor.actor_id)::text) > Filter: ((role_class_code)::text = > 'Judge'::text) > -> Index Scan using event_actor_speed on event_actor > (cost=0.00..695.15 rows=1275 width=73) > Index Cond: ((event_actor.actor_id)::text = > ("outer".actor_id)::text) > -> Index Scan using event_pkey on event (cost=0.00..5.83 > rows=1 width=52) > Index Cond: (("outer".event_id)::text = > (event.event_id)::text) > Filter: (event_date_time > '20021126'::bpchar) > > > -- > "You can't make a program without broken egos." > -- > Jeremiah Jahn > > > ---------------------------(end of broadcast)--------------------------- > TIP 3: if posting/reading through Usenet, please send an appropriate > subscribe-nomail command to majordomo@postgresql.org so that your > message can get through to the mailing list cleanly > From pgsql-performance-owner@postgresql.org Wed Nov 26 18:33:27 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 78F9AD1B442 for ; Wed, 26 Nov 2003 22:33:17 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 88758-05 for ; Wed, 26 Nov 2003 18:32:54 -0400 (AST) Received: from fuji.krosing.net (217-159-136-226-dsl.kt.estpak.ee [217.159.136.226]) by svr1.postgresql.org (Postfix) with ESMTP id D6412D1B494 for ; Wed, 26 Nov 2003 18:32:45 -0400 (AST) Received: from fuji.krosing.net (localhost.localdomain [127.0.0.1]) by fuji.krosing.net (8.12.8/8.12.8) with ESMTP id hAQMWULX003331; Thu, 27 Nov 2003 00:32:31 +0200 Received: (from hannu@localhost) by fuji.krosing.net (8.12.8/8.12.8/Submit) id hAQMWUrl003329; Thu, 27 Nov 2003 00:32:30 +0200 X-Authentication-Warning: fuji.krosing.net: hannu set sender to hannu@tm.ee using -f Subject: Re: cross table indexes or something? From: Hannu Krosing To: Jeremiah Jahn Cc: postgres performance In-Reply-To: <1069877651.22346.13.camel@bluejay.goodinassociates.com> References: <1069877651.22346.13.camel@bluejay.goodinassociates.com> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <1069885949.3176.57.camel@fuji.krosing.net> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 Date: Thu, 27 Nov 2003 00:32:30 +0200 X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/281 X-Sequence-Number: 4841 Jeremiah Jahn kirjutas K, 26.11.2003 kell 22:14: > I was wondering if there is something I can do that would act similar to > a index over more than one table. > > I have about 3 million people in my DB at the moment, they all have > roles, and many of them have more than one name. > > for example, a Judge will only have one name, but a Litigant could have > multiple aliases. Things go far to slow when I do a query on a judge > named smith. If you dont need all the judges named smith you could try to use LIMIT. Have you run ANALYZE ? Why does DB think that there is only one judge with name like SMITH% ? ------------- Hannu P.S. Always send EXPLAIN ANALYZE output if asking for advice on [PERFORM] ------------- Hannu From pgsql-performance-owner@postgresql.org Wed Nov 26 18:43:43 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 87083D1B494 for ; Wed, 26 Nov 2003 22:43:41 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 91445-01 for ; Wed, 26 Nov 2003 18:43:18 -0400 (AST) Received: from jhuml1.jhmi.edu (jhuml1.jhmi.edu [162.129.234.20]) by svr1.postgresql.org (Postfix) with ESMTP id 19E40D1B489 for ; Wed, 26 Nov 2003 18:43:10 -0400 (AST) Received: from jhuml1.jhmi.edu (jhuml1.jhmi.edu [162.129.234.20]) by jhuml1.jhmi.edu (PMDF V6.2-X17 #30839) with SMTP id <0HOZ006IGDS1HD@jhuml1.jhmi.edu> for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 17:43:16 -0500 (EST) Received: from jhuml1.jhmi.edu ([162.129.234.20]) by jhuml1.jhmi.edu (SAVSMTP 3.1.0.29) with SMTP id M2003112617431615949 for ; Wed, 26 Nov 2003 17:43:16 -0500 Received: from jhmimail.jhmi.edu (jhem2.jhmi.edu [162.129.8.23]) by jhuml1.jhmi.edu (PMDF V6.2-X17 #30839) id <0HOZ00801DS4MO@jhuml1.jhmi.edu> (original mail from lshao2@jhmi.edu) for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 17:43:16 -0500 (EST) Received: from jhmimail.jhmi.edu (jhem2.jhmi.edu [162.129.8.23]) by jhuml1.jhmi.edu (PMDF V6.2-X17 #30839) with ESMTP id <0HOZ0040TDS3GN@jhuml1.jhmi.edu> for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 17:43:16 -0500 (EST) Received: from [162.129.178.60] by jhmimail.jhmi.edu (mshttpd); Wed, 26 Nov 2003 22:46:24 +0000 (GMT) Date: Wed, 26 Nov 2003 22:46:24 +0000 (GMT) From: LIANHE SHAO Subject: Re: very large db performance question To: Neil Conway Cc: pgsql-performance@postgresql.org Message-id: <4e2c4d4e061f.4e061f4e2c4d@jhmimail.jhmi.edu> MIME-version: 1.0 X-Mailer: iPlanet Messenger Express 5.2 HotFix 1.17 (built Jun 23 2003) Content-type: text/plain; charset=us-ascii Content-language: en Content-transfer-encoding: 7bit Content-disposition: inline X-Accept-Language: en X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/282 X-Sequence-Number: 4842 Thanks for reply. Actually our database only supply some scientists to use (we predict that). so there is no workload problem. there is only very infrequent updates. the query is not complex. the problem is, we have one table that store most of the data ( with 200 million rows). In this table, there is a text column which we need to do full text search for each row. The result will then join the data from another table which has 30,000 rows. Now the query runs almost forever. I tried a small table with 2 million rows using the following simple command, it takes me about 6 seconds to get the result back. So, I get confused. That is why I ask: Is it the hardware problem or something else. (I just vacuumed the whole database yesterday). PGA=> select count (*) from expressiondata ; count --------- 2197497 (1 row) PGA=> explain select count (*) from expressiondata ; QUERY PLAN ------------------------------------------------------------------------------ Aggregate (cost=46731.71..46731.71 rows=1 width=0) -> Seq Scan on expressiondata (cost=0.00..41237.97 rows=2197497 width=0) (2 rows) Regards, William ----- Original Message ----- From: Neil Conway Date: Wednesday, November 26, 2003 10:03 pm Subject: Re: [PERFORM] very large db performance question > LIANHE SHAO writes: > > We will have a very large database to store microarray data (may > > exceed 80-100G some day). now we have 1G RAM, 2G Hz Pentium 4, 1 > > CPU. and enough hard disk. > > > Could anybody tell me that our hardware is an issue or not? > > IMHO the size of the DB is less relevant than the query workload. For > example, if you're storying 100GB of data but only doing a single > index scan on it every 10 seconds, any modern machine with enough HD > space should be fine. > > If you give us an idea of the # of queries you expect per second, the > approximate mix of reads and writes, and some idea of how complex the > queries are, we might be able to give you some better advice. > > -Neil > > > From pgsql-performance-owner@postgresql.org Wed Nov 26 20:09:53 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id C8FC7D1B492 for ; Thu, 27 Nov 2003 00:09:49 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 97927-04 for ; Wed, 26 Nov 2003 20:09:27 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id D98CED1B45C for ; Wed, 26 Nov 2003 20:09:18 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hAR09P19018734; Wed, 26 Nov 2003 19:09:25 -0500 (EST) To: Roger Ging Cc: pgsql-performance@postgresql.org Subject: Re: Followup - expression (functional) index use in joins In-reply-to: <3FC51B2C.2030700@paccomsys.com> References: <3FC4F360.2090609@paccomsys.com> <200311261912.01286.dev@archonet.com> <3FC51B2C.2030700@paccomsys.com> Comments: In-reply-to Roger Ging message dated "Wed, 26 Nov 2003 13:29:16 -0800" Date: Wed, 26 Nov 2003 19:09:25 -0500 Message-ID: <18733.1069891765@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/283 X-Sequence-Number: 4843 Roger Ging writes: > Ran vacuum analyse on both program and logfile tables. Estimates are > more in line with reality now, And they are what now? You really can't expect to get useful help here when you're being so miserly with the details ... FWIW, I suspect you could force 7.4 to generate 7.3's plan by setting enable_mergejoin to off (might have to also set enable_hashjoin to off, if it then tries for a hash join). 7.3 could not even consider those join types in this example, while 7.4 can. The interesting question from my perspective is why the planner is guessing wrong about the relative costs of the plans. EXPLAIN ANALYZE results with each type of join forced would be useful to look at. regards, tom lane From pgsql-performance-owner@postgresql.org Wed Nov 26 20:47:58 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 33278D1B457 for ; Thu, 27 Nov 2003 00:47:55 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 01406-06 for ; Wed, 26 Nov 2003 20:47:32 -0400 (AST) Received: from houston.familyhealth.com.au (fhnet.arach.net.au [203.22.197.21]) by svr1.postgresql.org (Postfix) with ESMTP id 032C0D1B43F for ; Wed, 26 Nov 2003 20:47:22 -0400 (AST) Received: from familyhealth.com.au (work-40.internal [192.168.0.40]) by houston.familyhealth.com.au (8.12.9p1/8.12.9) with ESMTP id hAR0l9oD051233; Thu, 27 Nov 2003 08:47:10 +0800 (WST) (envelope-from chriskl@familyhealth.com.au) Message-ID: <3FC54A82.2040605@familyhealth.com.au> Date: Thu, 27 Nov 2003 08:51:14 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: LIANHE SHAO Cc: pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> In-Reply-To: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/284 X-Sequence-Number: 4844 > Which one is better (performance/easier to use), > tsearch2 or fulltextindex? > there is an example how to use fulltextindex in the > techdocs, but I checked the contrib/fulltextindex > package, there is a WARNING that fulltextindex is > much slower than tsearch2. but tsearch2 seems > complex to use, and I can not find a good example. > Which one I should use? Any suggestions? I believe I wrote that warning :) Tsearch2 is what you should use. Yes, it's more complicated but it's HEAPS faster and seriously powerful. Just read the README file. You could also try out the original tsearch (V1), but that will probably be superceded soon, now that tsearch2 is around. Chris From pgsql-performance-owner@postgresql.org Wed Nov 26 20:59:58 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 8E2D3D1B439 for ; Thu, 27 Nov 2003 00:59:35 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 04518-01 for ; Wed, 26 Nov 2003 20:59:13 -0400 (AST) Received: from houston.familyhealth.com.au (fhnet.arach.net.au [203.22.197.21]) by svr1.postgresql.org (Postfix) with ESMTP id C9C45D1B43F for ; Wed, 26 Nov 2003 20:59:03 -0400 (AST) Received: from familyhealth.com.au (work-40.internal [192.168.0.40]) by houston.familyhealth.com.au (8.12.9p1/8.12.9) with ESMTP id hAR0x8oD051309; Thu, 27 Nov 2003 08:59:09 +0800 (WST) (envelope-from chriskl@familyhealth.com.au) Message-ID: <3FC54D52.4040207@familyhealth.com.au> Date: Thu, 27 Nov 2003 09:03:14 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: LIANHE SHAO Cc: Neil Conway , pgsql-performance@postgresql.org Subject: Re: very large db performance question References: <4e2c4d4e061f.4e061f4e2c4d@jhmimail.jhmi.edu> In-Reply-To: <4e2c4d4e061f.4e061f4e2c4d@jhmimail.jhmi.edu> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/285 X-Sequence-Number: 4845 > Thanks for reply. Actually our database only supply > some scientists to use (we predict that). so there > is no workload problem. there is only very > infrequent updates. the query is not complex. the > problem is, we have one table that store most of the > data ( with 200 million rows). In this table, there > is a text column which we need to do full text > search for each row. The result will then join the > data from another table which has 30,000 rows. Now > the query runs almost forever. Use TSearch2. > I tried a small table with 2 million rows using the > following simple command, it takes me about 6 > seconds to get the result back. So, I get confused. > That is why I ask: Is it the hardware problem or > something else. (I just vacuumed the whole database > yesterday). > > PGA=> select count (*) from expressiondata ; > count > --------- > 2197497 > (1 row) select count(*) on a postgres table ALWAYS does a sequential scan. Just don't do it. There are technical reasons (MVCC) why this is so. It's a bad "test". Chris From pgsql-performance-owner@postgresql.org Wed Nov 26 21:04:37 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EE9C5D1B49C for ; Thu, 27 Nov 2003 01:04:31 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 03659-02 for ; Wed, 26 Nov 2003 21:04:09 -0400 (AST) Received: from rlx13.zapatec.com (66-117-144-213.zapatec.lmi.net [66.117.144.213]) by svr1.postgresql.org (Postfix) with ESMTP id E663BD1B448 for ; Wed, 26 Nov 2003 21:03:59 -0400 (AST) Received: from rlx11.zapatec.com (rlx11.pr.zapatec.com [192.168.1.132]) by rlx13.zapatec.com (Postfix) with ESMTP id 8A4F6A941 for ; Wed, 26 Nov 2003 17:03:53 -0800 (PST) Received: (from dror@localhost) by rlx11.zapatec.com (8.12.3/8.12.3/Submit) id hAR13qsa044103 for pgsql-performance@postgresql.org; Wed, 26 Nov 2003 17:03:52 -0800 (PST) (envelope-from dror) Date: Wed, 26 Nov 2003 17:03:52 -0800 From: Dror Matalon To: pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or Message-ID: <20031127010352.GT30893@rlx11.zapatec.com> References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> <3FC54A82.2040605@familyhealth.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3FC54A82.2040605@familyhealth.com.au> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/286 X-Sequence-Number: 4846 On Thu, Nov 27, 2003 at 08:51:14AM +0800, Christopher Kings-Lynne wrote: > >Which one is better (performance/easier to use), > >tsearch2 or fulltextindex? > >there is an example how to use fulltextindex in the > >techdocs, but I checked the contrib/fulltextindex > >package, there is a WARNING that fulltextindex is > >much slower than tsearch2. but tsearch2 seems > >complex to use, and I can not find a good example. > >Which one I should use? Any suggestions? > > I believe I wrote that warning :) > > Tsearch2 is what you should use. Yes, it's more complicated but it's > HEAPS faster and seriously powerful. > Can you provide some numbers please, both for creating full text indexes as well as for searching them? I tried to use tsearch and it seemed like just creating a full text index on million+ records took forever. > Just read the README file. > > You could also try out the original tsearch (V1), but that will probably > be superceded soon, now that tsearch2 is around. > > Chris > > > > ---------------------------(end of broadcast)--------------------------- > TIP 5: Have you checked our extensive FAQ? > > http://www.postgresql.org/docs/faqs/FAQ.html -- Dror Matalon Zapatec Inc 1700 MLK Way Berkeley, CA 94709 http://www.fastbuzz.com http://www.zapatec.com From pgsql-performance-owner@postgresql.org Wed Nov 26 23:35:10 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 1A7C3D1B47C for ; Thu, 27 Nov 2003 03:33:32 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 20533-03 for ; Wed, 26 Nov 2003 23:33:10 -0400 (AST) Received: from gp.word-to-the-wise.com (gp.word-to-the-wise.com [64.71.176.18]) by svr1.postgresql.org (Postfix) with ESMTP id 2556CD1B4CE for ; Wed, 26 Nov 2003 23:33:00 -0400 (AST) Received: by gp.word-to-the-wise.com (Postfix, from userid 500) id EBB9990000D; Wed, 26 Nov 2003 19:28:31 -0800 (PST) Date: Wed, 26 Nov 2003 19:28:31 -0800 From: Steve Atkins To: pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or Message-ID: <20031127032831.GA4836@gp.word-to-the-wise.com> References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> <3FC54A82.2040605@familyhealth.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3FC54A82.2040605@familyhealth.com.au> User-Agent: Mutt/1.4.1i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/287 X-Sequence-Number: 4847 On Thu, Nov 27, 2003 at 08:51:14AM +0800, Christopher Kings-Lynne wrote: > >Which one is better (performance/easier to use), > >tsearch2 or fulltextindex? > >there is an example how to use fulltextindex in the > >techdocs, but I checked the contrib/fulltextindex > >package, there is a WARNING that fulltextindex is > >much slower than tsearch2. but tsearch2 seems > >complex to use, and I can not find a good example. > >Which one I should use? Any suggestions? > > I believe I wrote that warning :) > > Tsearch2 is what you should use. Yes, it's more complicated but it's > HEAPS faster and seriously powerful. Does anyone have any metrics on how fast tsearch2 actually is? I tried it on a synthetic dataset of a million documents of a hundred words each and while insertions were impressively fast I gave up on the search after 10 minutes. Broken? Unusable slow? This was on the last 7.4 release candidate. Cheers, Steve From pgsql-performance-owner@postgresql.org Thu Nov 27 00:39:36 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 062B2D1B461 for ; Thu, 27 Nov 2003 04:38:59 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 24019-09 for ; Thu, 27 Nov 2003 00:38:20 -0400 (AST) Received: from houston.familyhealth.com.au (fhnet.arach.net.au [203.22.197.21]) by svr1.postgresql.org (Postfix) with ESMTP id EA5D8D1B45F for ; Thu, 27 Nov 2003 00:38:03 -0400 (AST) Received: from familyhealth.com.au (work-40.internal [192.168.0.40]) by houston.familyhealth.com.au (8.12.9p1/8.12.9) with ESMTP id hAR4beoD053430; Thu, 27 Nov 2003 12:37:41 +0800 (WST) (envelope-from chriskl@familyhealth.com.au) Message-ID: <3FC58097.8090803@familyhealth.com.au> Date: Thu, 27 Nov 2003 12:41:59 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Steve Atkins Cc: pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> <3FC54A82.2040605@familyhealth.com.au> <20031127032831.GA4836@gp.word-to-the-wise.com> In-Reply-To: <20031127032831.GA4836@gp.word-to-the-wise.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/288 X-Sequence-Number: 4848 > Does anyone have any metrics on how fast tsearch2 actually is? > > I tried it on a synthetic dataset of a million documents of a hundred > words each and while insertions were impressively fast I gave up on > the search after 10 minutes. > > Broken? Unusable slow? This was on the last 7.4 release candidate. I just created a 1.1million row dataset by copying one of our 30000 row production tables and just taking out the txtidx column. Then I inserted it into itself until it had 1.1 million rows. Then I created the GiST index - THAT took forever - seriously like 20 mins or half an hour or something. Now, to find a word: select * from tsearchtest where ftiidx ## 'curry'; Time: 9760.75 ms The AND of two words: Time: 103.61 ms The AND of three words: select * from tsearchtest where ftiidx ## 'curry&green&thai'; Time: 61.86 ms And now a one word query now that buffers are cached: select * from tsearchtest where ftiidx ## 'curry'; Time: 444.89 ms So, I have no idea why you think it's slow? Perhaps you forgot the 'create index using gist' step? Also, if you use the NOT (!) operand, you can get yourself into a really slow situation. Chris From pgsql-performance-owner@postgresql.org Thu Nov 27 01:17:49 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id A9967D1B43C for ; Thu, 27 Nov 2003 05:17:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 25728-07 for ; Thu, 27 Nov 2003 01:17:09 -0400 (AST) Received: from gp.word-to-the-wise.com (gp.word-to-the-wise.com [64.71.176.18]) by svr1.postgresql.org (Postfix) with ESMTP id 7FA97D1B441 for ; Thu, 27 Nov 2003 01:17:08 -0400 (AST) Received: by gp.word-to-the-wise.com (Postfix, from userid 500) id 195A190000D; Wed, 26 Nov 2003 21:12:30 -0800 (PST) Date: Wed, 26 Nov 2003 21:12:30 -0800 From: Steve Atkins To: pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or Message-ID: <20031127051230.GA6212@gp.word-to-the-wise.com> References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> <3FC54A82.2040605@familyhealth.com.au> <20031127032831.GA4836@gp.word-to-the-wise.com> <3FC58097.8090803@familyhealth.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3FC58097.8090803@familyhealth.com.au> User-Agent: Mutt/1.4.1i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/289 X-Sequence-Number: 4849 On Thu, Nov 27, 2003 at 12:41:59PM +0800, Christopher Kings-Lynne wrote: > >Does anyone have any metrics on how fast tsearch2 actually is? > > > >I tried it on a synthetic dataset of a million documents of a hundred > >words each and while insertions were impressively fast I gave up on > >the search after 10 minutes. > > > >Broken? Unusable slow? This was on the last 7.4 release candidate. > > I just created a 1.1million row dataset by copying one of our 30000 row > production tables and just taking out the txtidx column. Then I > inserted it into itself until it had 1.1 million rows. > > Then I created the GiST index - THAT took forever - seriously like 20 > mins or half an hour or something. > > Now, to find a word: > > select * from tsearchtest where ftiidx ## 'curry'; > Time: 9760.75 ms > So, I have no idea why you think it's slow? Perhaps you forgot the > 'create index using gist' step? No, it was indexed. Thanks, that was the datapoint I was looking for. It _can_ run fast, so I just need to work out what's going on. (It's hard to diagnose a slow query when you've no idea whether it's really 'slow'). Cheers, Steve From pgsql-performance-owner@postgresql.org Thu Nov 27 04:44:51 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id B9C41D1B451 for ; Thu, 27 Nov 2003 08:41:57 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 47479-07 for ; Thu, 27 Nov 2003 04:41:28 -0400 (AST) Received: from ecbull20.frec.bull.fr (ecbull20.frec.bull.fr [129.183.4.3]) by svr1.postgresql.org (Postfix) with ESMTP id BE283D1B43F for ; Thu, 27 Nov 2003 04:41:25 -0400 (AST) Received: from ecn002.frec.bull.fr (ecn002.frec.bull.fr [129.183.4.6]) by ecbull20.frec.bull.fr (8.9.2/8.9.1) with ESMTP id JAA20544; Thu, 27 Nov 2003 09:42:29 +0100 Received: from BULL.NET ([129.183.148.134]) by ecn002.frec.bull.fr (Lotus Domino Release 5.0.12) with ESMTP id 2003112709454775:638 ; Thu, 27 Nov 2003 09:45:47 +0100 Message-ID: <3FC5B8F6.49A663E7@BULL.NET> Date: Thu, 27 Nov 2003 09:42:30 +0100 From: Thierry Missimilly Organization: BSIS/R&D X-Mailer: Mozilla 4.75 [en] (Windows NT 5.0; U) X-Accept-Language: fr,en MIME-Version: 1.0 To: LIANHE SHAO Cc: pgsql-performance@postgresql.org Subject: Re: very large db performance question References: <4cbacd4cb3c2.4cb3c24cbacd@jhmimail.jhmi.edu> X-MIMETrack: Itemize by SMTP Server on ECN002/FR/BULL(Release 5.0.12 |February 13, 2003) at 27/11/2003 09:45:47, Serialize by Router on ECN002/FR/BULL(Release 5.0.12 |February 13, 2003) at 27/11/2003 09:45:55, Serialize complete at 27/11/2003 09:45:55 Content-Type: multipart/mixed; boundary="------------F715D175693138A496FD42FA" X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/290 X-Sequence-Number: 4850 This is a multi-part message in MIME format. --------------F715D175693138A496FD42FA Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi, I have done some performance tests on 1Gb and 4 Gb Databases on a mono Pentium 4 , 1 Gb RAM, IDE disk, SCSI disks and RAID0 LUN on DAS 5300 on Linux RedHat 7.3. In each cases my tests make select, update and insert. One of them is pgbench. You can find it in Postgres/contrib/pgbench. The other one is DBT1 form OSDL. I have port it on Postgres and you can find it in Source Forge. Jenny Wang is writting a better Postgres DBT1 based on C transactions instead of PL/pgSQL Transactions. With this size of database, even after a fine tuning of Postgres the problem is I/O Wait. So in you case with 100 Gb, you will have I/O Wait. To resume my observations regarding diskss performances : 1) IDE disk are the slower. 2) SCSI disks are a little more faster but you can decrease I/O Wait by 25% by creating a stripped volume group on 3 disks. 3) A RAID0 on 5 DAS5300 disks improve again performances by 20% as the DAS Storage Processeur use internal caches One thing, very important in my case was the time of (hot) backup / restore. In that case the pgbench database schema is to simple to have an idea but DBT1 schema is enough complex and on the RAID0 LUN the backup takes 12 min but the restore takes 16 min + 10 min to recreate the indexes + 255 min to recreate the Foreign Keys. So 4h41 for a 4Gb database. That means for a 100 Gb database, if your schema as Foreign keys and indexes : about 5 hours to backup and 117 hours to restore (~5 days). So, if your database in a critical database, it is better to use cold backup with Snapshot tools. Regards, Thierry Missimilly LIANHE SHAO wrote: > Hello All, > We will have a very large database to store > microarray data (may exceed 80-100G some day). now > we have 1G RAM, 2G Hz Pentium 4, 1 CPU. and enough > hard disk. > > I never touched such large database before. I ask > several dbas if the hardware is ok, some said it is > ok for the query, but I am not so convinced. Because > I check the mailing list and learned that it is not > unreasonable to take several minutes to do the > query. But I want to query to be as fast as possible. > > Could anybody tell me that our hardware is an issue > or not? do we really need better hardware to make > real difference? > > Regards, > William > > ---------------------------(end of broadcast)--------------------------- > TIP 8: explain analyze is your friend --------------F715D175693138A496FD42FA Content-Type: text/x-vcard; charset=us-ascii; name="THIERRY.MISSIMILLY.vcf" Content-Description: Card for Thierry Missimilly Content-Disposition: attachment; filename="THIERRY.MISSIMILLY.vcf" Content-Transfer-Encoding: 7bit begin:vcard n:Missimilly;Thierry tel;fax:+33 (0)4 76 29 78 78 tel;work:+33 (0)4 76 29 74 54 x-mozilla-html:FALSE url:http:\\www.bull.com org:BIS/R&D adr:;;Bull SA, 1, rue de provence - BP 208;ECHIROLLES;;38432;FRANCE version:2.1 email;internet:Thierry.Missimilly@bull.net x-mozilla-cpt:;-18184 fn:Thierry Missimilly end:vcard --------------F715D175693138A496FD42FA-- From pgsql-performance-owner@postgresql.org Thu Nov 27 15:58:44 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 18B76D1B47A for ; Thu, 27 Nov 2003 19:58:40 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 36366-03 for ; Thu, 27 Nov 2003 15:58:10 -0400 (AST) Received: from smtp3.mail.be.easynet.net (bigglesworth.mail.be.easynet.net [212.100.160.67]) by svr1.postgresql.org (Postfix) with ESMTP id A2E71D1B450 for ; Thu, 27 Nov 2003 15:58:08 -0400 (AST) Received: from 213-193-176-107.adsl.easynet.be ([213.193.176.107]) by smtp3.mail.be.easynet.net with esmtp (Exim 4.22) id 1APSH0-0003xL-Pw for pgsql-performance@postgresql.org; Thu, 27 Nov 2003 20:57:58 +0100 From: Stefan Champailler Reply-To: schampailler@easynet.be To: pgsql-performance@postgresql.org Subject: Re: Impossibly slow DELETEs Date: Thu, 27 Nov 2003 20:57:06 +0100 User-Agent: KMail/1.5.3 References: <200311252256.56433.schampailler@easynet.be> <3FC3E984.3020508@potentialtech.com> In-Reply-To: <3FC3E984.3020508@potentialtech.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200311272057.06718.schampailler@easynet.be> X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/291 X-Sequence-Number: 4851 I did not conduct much more test but from what I've seen, it looks like the ODBC driver is in the doldrums, not PG. For example, when I run my software on Windows rather than Linux, everything just works as expected. Sorry for disturbing. And btw, I use ODBC because my target DB is Oracle and I've been requested to access it throguh ODBC. So, because I don't have Oracle, I do most of my development with PG and then I'll port to Oracle. Since I'm doing "simple" stuff, PG is almost 100% compatible with Oracle. (and before you ask, no, they don't give me the proper dev environment, bastards :)) Thanks for all the answers. Stefan > Stefan Champailler wrote: > > Dear You all, > > > > (please tell me if this has already been discussed, I was unable to find > > any convincing information) > > > > I'm developing a small application, tied to a PG 7.4 beta 5 (i didn't > > upgrade). The DB i use is roughly 20 tales each of them containing at > > most 30 records (I'm still in development). I can provide a whole dump if > > necessary. I access the DB throug IODBC (Suse Linux 8.1), through PHP. > > The machine everything runs on is 512M of Ram, 2.5GHz speed. So I assume > > it should be blazingly fast. > > > > So here's my trouble : some DELETE statement take up to 1 minute to > > complete (but not always, sometimes it's fast, sometimes it's that slow). > > Here's a typical one : DELETE FROM response_bool WHERE response_id = > > '125' The response_bool table has no foreing key and no index on > > response_id column. No foreign key reference the response_bool table. > > There are 6 rows in the table (given that size, I assumed that an index > > was not necessary). > > > > So 1 minute to complete look like I did something REALLY bad. > > > > It is my feeling that doing the same query with psql works without > > problem, but I can't be sure. > > I think that last sentence is the crux of the problem. If you can > establish for sure that the unreasonable delay is _only_ there when the > command is issued through IODBC, then it's not a Postgresql problem. > > Out of curiosity, why are you using ODBC for PHP anyway? PHP has > Postgresql libraries that work very well. I use them quite often without > problems. From pgsql-performance-owner@postgresql.org Fri Nov 28 01:09:36 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EA2D0D1B46F for ; Fri, 28 Nov 2003 05:09:29 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 82788-07 for ; Fri, 28 Nov 2003 01:08:59 -0400 (AST) Received: from gp.word-to-the-wise.com (gp.word-to-the-wise.com [64.71.176.18]) by svr1.postgresql.org (Postfix) with ESMTP id 30CE2D1B468 for ; Fri, 28 Nov 2003 01:08:58 -0400 (AST) Received: by gp.word-to-the-wise.com (Postfix, from userid 500) id C273C90000D; Thu, 27 Nov 2003 21:04:17 -0800 (PST) Date: Thu, 27 Nov 2003 21:04:17 -0800 From: Steve Atkins To: pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or Message-ID: <20031128050417.GA14227@gp.word-to-the-wise.com> References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> <3FC54A82.2040605@familyhealth.com.au> <20031127032831.GA4836@gp.word-to-the-wise.com> <3FC58097.8090803@familyhealth.com.au> <20031127051230.GA6212@gp.word-to-the-wise.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20031127051230.GA6212@gp.word-to-the-wise.com> User-Agent: Mutt/1.4.1i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/292 X-Sequence-Number: 4852 On Wed, Nov 26, 2003 at 09:12:30PM -0800, Steve Atkins wrote: > On Thu, Nov 27, 2003 at 12:41:59PM +0800, Christopher Kings-Lynne wrote: > > >Does anyone have any metrics on how fast tsearch2 actually is? > > > > > >I tried it on a synthetic dataset of a million documents of a hundred > > >words each and while insertions were impressively fast I gave up on > > >the search after 10 minutes. > > > > > >Broken? Unusable slow? This was on the last 7.4 release candidate. > > > > I just created a 1.1million row dataset by copying one of our 30000 row > > production tables and just taking out the txtidx column. Then I > > inserted it into itself until it had 1.1 million rows. > > > > Then I created the GiST index - THAT took forever - seriously like 20 > > mins or half an hour or something. > > > > Now, to find a word: > > > > select * from tsearchtest where ftiidx ## 'curry'; > > Time: 9760.75 ms > > > So, I have no idea why you think it's slow? Perhaps you forgot the > > 'create index using gist' step? > > No, it was indexed. > > Thanks, that was the datapoint I was looking for. It _can_ run fast, so > I just need to work out what's going on. (It's hard to diagnose a slow > query when you've no idea whether it's really 'slow'). Looking at it further, something is very broken, possibly with GIST indices, possibly with tsearch2s use of 'em. This is on a newly built 7.4 installation, built with 64 bit datetimes, but completely stock other than that. Stock gcc 3.3.2, Linux, somewhat elderly 2.4.18 kernel. Running on a 1.5GHz single processor Athlon with a half gig of RAM. Configuration set to use 20% of RAM as shared buffers (amongst other settings, this was the last of a range I tried looking for variation). Software RAID0 across two 7200RPM SCSI drives, reiserfs (it's a development box, not a production system). System completely idle apart from postgresql. 269000 rows, each row having 400 words. Analyzed. Running the select query given below appears to pause a process trying to insert into the table completely (locking issue? I/O bandwidth?). top shows the select below consuming <2% of CPU and iostat shows it reading ~2800 blocks/second from each of the two RAID drives. Physical size of the database is under 3 gigs, including toast and index tables. The select query takes around 6 minutes (consistently, even if the same identical query is repeated). For entertainment, I turned off indexscan and the query takes 1 minute with a simple seqscan. Any thoughts? Cheers, Steve => select count(*) from ftstest; count -------- 269000 (1 row) => \d ftstest Table "public.ftstest" Column | Type | Modifiers --------+----------+---------------------------------------------------------- idx | integer | not null default nextval('public.ftstest_idx_seq'::text) words | text | not null idxfti | tsvector | not null Indexes: "ftstest_idx" gist (idxfti) => explain analyze select idx from ftstest where idxfti @@ 'dominican'::tsquery; QUERY PLAN ----------------------------------------------------------------------------------------------------------------------------------- Index Scan using ftstest_idx on ftstest (cost=0.00..515.90 rows=271 width=4) (actual time=219.694..376042.428 rows=4796 loops=1) Index Cond: (idxfti @@ '\'dominican\''::tsquery) Filter: (idxfti @@ '\'dominican\''::tsquery) Total runtime: 376061.541 ms (4 rows) ((Set enable_indexscan=false)) => explain analyze select idx from ftstest where idxfti @@ 'dominican'::tsquery; QUERY PLAN -------------------------------------------------------------------------------------------------------------- Seq Scan on ftstest (cost=0.00..5765.88 rows=271 width=4) (actual time=42.589..62158.285 rows=4796 loops=1) Filter: (idxfti @@ '\'dominican\''::tsquery) Total runtime: 62182.277 ms (3 rows) From pgsql-performance-owner@postgresql.org Fri Nov 28 01:14:59 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id AB67ED1B44D for ; Fri, 28 Nov 2003 05:14:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 86782-02 for ; Fri, 28 Nov 2003 01:14:04 -0400 (AST) Received: from houston.familyhealth.com.au (fhnet.arach.net.au [203.22.197.21]) by svr1.postgresql.org (Postfix) with ESMTP id 6057BD1B446 for ; Fri, 28 Nov 2003 01:14:02 -0400 (AST) Received: from familyhealth.com.au (work-40.internal [192.168.0.40]) by houston.familyhealth.com.au (8.12.9p1/8.12.9) with ESMTP id hAS5DfoD062128; Fri, 28 Nov 2003 13:13:41 +0800 (WST) (envelope-from chriskl@familyhealth.com.au) Message-ID: <3FC6DAB8.4080106@familyhealth.com.au> Date: Fri, 28 Nov 2003 13:18:48 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Steve Atkins Cc: pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> <3FC54A82.2040605@familyhealth.com.au> <20031127032831.GA4836@gp.word-to-the-wise.com> <3FC58097.8090803@familyhealth.com.au> <20031127051230.GA6212@gp.word-to-the-wise.com> <20031128050417.GA14227@gp.word-to-the-wise.com> In-Reply-To: <20031128050417.GA14227@gp.word-to-the-wise.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/293 X-Sequence-Number: 4853 > Any thoughts? Actually, I ran my tests using tsearch V1. I wonder if there has been some weird regression between tsearch 1 and 2? hris From pgsql-performance-owner@postgresql.org Fri Nov 28 01:23:02 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id F27FAD1B481 for ; Fri, 28 Nov 2003 05:22:53 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 89056-01 for ; Fri, 28 Nov 2003 01:22:22 -0400 (AST) Received: from houston.familyhealth.com.au (fhnet.arach.net.au [203.22.197.21]) by svr1.postgresql.org (Postfix) with ESMTP id 5C006D1B45E for ; Fri, 28 Nov 2003 01:22:21 -0400 (AST) Received: from familyhealth.com.au (work-40.internal [192.168.0.40]) by houston.familyhealth.com.au (8.12.9p1/8.12.9) with ESMTP id hAS5LioD062170; Fri, 28 Nov 2003 13:21:45 +0800 (WST) (envelope-from chriskl@familyhealth.com.au) Message-ID: <3FC6DC9B.4060904@familyhealth.com.au> Date: Fri, 28 Nov 2003 13:26:51 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Christopher Kings-Lynne Cc: Steve Atkins , pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> <3FC54A82.2040605@familyhealth.com.au> <20031127032831.GA4836@gp.word-to-the-wise.com> <3FC58097.8090803@familyhealth.com.au> <20031127051230.GA6212@gp.word-to-the-wise.com> <20031128050417.GA14227@gp.word-to-the-wise.com> <3FC6DAB8.4080106@familyhealth.com.au> In-Reply-To: <3FC6DAB8.4080106@familyhealth.com.au> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/294 X-Sequence-Number: 4854 >> Any thoughts? > > > Actually, I ran my tests using tsearch V1. I wonder if there has been > some weird regression between tsearch 1 and 2? I also ran my tests on 7.3.4 :( Chris From pgsql-performance-owner@postgresql.org Fri Nov 28 01:59:05 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 11AE0D1B493 for ; Fri, 28 Nov 2003 05:58:59 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 78976-09 for ; Fri, 28 Nov 2003 01:58:28 -0400 (AST) Received: from mail.delegated.net (nitro.ezprovider.net [66.199.148.100]) by svr1.postgresql.org (Postfix) with SMTP id A91E2D1B45A for ; Fri, 28 Nov 2003 01:58:11 -0400 (AST) Received: (qmail 32296 invoked from network); 28 Nov 2003 05:57:55 -0000 Received: from h24-85-181-182.vn.shawcable.net (HELO delegated.net) (24.85.181.182) by nitro.ezprovider.net with SMTP; 28 Nov 2003 05:57:55 -0000 Message-ID: <3FC6E405.5010509@delegated.net> Date: Thu, 27 Nov 2003 21:58:29 -0800 From: Jonathan Knopp User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.6a) Gecko/20031030 X-Accept-Language: en-us, en MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: simple left join slows query more than expected Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/295 X-Sequence-Number: 4855 I'v spent a couple days playing with this problem and searching the mailing lists and docs etc but come up with nothing. Any help would be much appreciated. Setup is postgres 7.3.2 on redhat 7.1 on a 1.3GHz Athlon machine with 1G pc133 ram and SCSI. Here is the same query with the addition of a left join onto a list of contacts to grab the last name of each connected contact. I'd think this should be real quick since it jsut has to grab around 100 names from the list, and if its smart enough to grab just distinct IDs, then it's just like 10 rows it has to grab using the primary field. But as far as i can tell (and i may VERY well be reading the explain syntax wrong), it is grabbing them all and joining them first, rather than doing the operation that limits the result rows to a mere 100 and THEN doing the join to contacts. It would be faster if i did a separate query using a big IN(id1,id2,...) condition, which makes no sense to me. Plus i REALLY want to avoid this as the selected fields and the joins and conditions are all variable and controlled (indirectly and transparently) by the user. Point is, why does a simple left join slow things down so much? in my experience (primarily with mysql but also over a year with postgre) simple left joins are usually quite quick. I can only guess that a bad plan is being chosen. PLEASE don't tell me i need to store a copy of the names in the events table to get acceptable speed, cause this would be plain sacrilegious in terms of DB design. Or is this simply as fast as these queries can go? Just seems too long for the work that's being done IME. events table has 12355 rows contacts has 20064 event_managers has 8502 All fields with conditions (object_ids, contact, event_id, user_id, deleted_on) are indexed with btree. Here is the query with the left join. sauce=# explain analyze SELECT top.object_id , top.who, top.datetime, top.priority, top.subject, top.action, top_contact_.last_name, top.object_id, top_contact_.object_id FROM event_managers AS managers JOIN ONLY events AS top ON(managers.event_id=top.object_id) LEFT JOIN contacts AS top_contact_ ON(top.contact=top_contact_.object_id and top_contact_.deleted_on IS NULL) WHERE true AND managers.user_id=238; QUERY PLAN -------------------------------------------------------------------------------------------------------------------------------------------- Merge Join (cost=5569.24..5671.22 rows=100 width=91) (actual time=485.95..526.25 rows=208 loops=1) Merge Cond: ("outer".contact = "inner".object_id) Join Filter: ("inner".deleted_on IS NULL) -> Sort (cost=2467.17..2467.42 rows=100 width=60) (actual time=143.67..143.75 rows=208 loops=1) Sort Key: top.contact -> Hash Join (cost=143.63..2463.83 rows=100 width=60) (actual time=0.89..142.64 rows=208 loops=1) Hash Cond: ("outer".object_id = "inner".event_id) -> Seq Scan on events top (cost=0.00..1830.19 rows=12219 width=56) (actual time=0.05..131.33 rows=12219 loops=1) -> Hash (cost=143.45..143.45 rows=69 width=4) (actual time=0.65..0.65 rows=0 loops=1) -> Index Scan using event_managers_user_id on event_managers managers (cost=0.00..143.45 rows=69 width=4) (actual time=0.14..0.50 rows=139 loops=1) Index Cond: (user_id = 238) -> Sort (cost=3102.07..3152.23 rows=20064 width=31) (actual time=342.23..360.29 rows=19964 loops=1) Sort Key: top_contact_.object_id -> Append (cost=0.00..1389.64 rows=20064 width=31) (actual time=0.06..115.63 rows=20064 loops=1) -> Seq Scan on contacts top_contact_ (cost=0.00..1383.43 rows=20043 width=31) (actual time=0.06..101.04 rows=20043 loops=1) -> Seq Scan on users top_contact_ (cost=0.00..6.21 rows=21 width=31) (actual time=0.05..0.29 rows=21 loops=1) Total runtime: 527.47 msec (17 rows) The same thing but without the left join. Much faster. Anything slower than this would be unacceptable, especailly given how small the tables are at this point. They are expected to grow ALOT bigger within a year. sauce=# explain analyze SELECT top.object_id , top.who, top.datetime, top.priority, top.subject, top.action, top.object_id FROM event_managers AS managers JOIN ONLY events AS top ON(managers.event_id=top.object_id) WHERE true AND managers.user_id=238; QUERY PLAN --------------------------------------------------------------------------------------------------------------------------------------------------------------- Hash Join (cost=143.63..2463.83 rows=100 width=56) (actual time=1.48..137.74 rows=208 loops=1) Hash Cond: ("outer".object_id = "inner".event_id) -> Seq Scan on events top (cost=0.00..1830.19 rows=12219 width=52) (actual time=0.06..125.80 rows=12219 loops=1) -> Hash (cost=143.45..143.45 rows=69 width=4) (actual time=1.20..1.20 rows=0 loops=1) -> Index Scan using event_managers_user_id on event_managers managers (cost=0.00..143.45 rows=69 width=4) (actual time=0.21..1.03 rows=139 loops=1) Index Cond: (user_id = 238) Total runtime: 137.96 msec (7 rows) again, many thanks for any feedback! From pgsql-performance-owner@postgresql.org Fri Nov 28 13:56:08 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 64A3CD1B433 for ; Fri, 28 Nov 2003 17:56:07 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 67667-10 for ; Fri, 28 Nov 2003 13:55:40 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id 71854D1B432 for ; Fri, 28 Nov 2003 13:55:39 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id hASHtc19029110; Fri, 28 Nov 2003 12:55:38 -0500 (EST) To: Jonathan Knopp Cc: pgsql-performance@postgresql.org Subject: Re: simple left join slows query more than expected In-reply-to: <3FC6E405.5010509@delegated.net> References: <3FC6E405.5010509@delegated.net> Comments: In-reply-to Jonathan Knopp message dated "Thu, 27 Nov 2003 21:58:29 -0800" Date: Fri, 28 Nov 2003 12:55:38 -0500 Message-ID: <29109.1070042138@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/296 X-Sequence-Number: 4856 Jonathan Knopp writes: > I'v spent a couple days playing with this problem and searching the mailing lists and > docs etc but come up with nothing. Any help would be much appreciated. Don't use inheritance to define the contacts table. regards, tom lane From pgsql-performance-owner@postgresql.org Fri Nov 28 16:58:29 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id EEE08D1B435 for ; Fri, 28 Nov 2003 20:42:12 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 95967-06 for ; Fri, 28 Nov 2003 16:41:43 -0400 (AST) Received: from gp.word-to-the-wise.com (gp.word-to-the-wise.com [64.71.176.18]) by svr1.postgresql.org (Postfix) with ESMTP id EBCACD1B438 for ; Fri, 28 Nov 2003 16:41:39 -0400 (AST) Received: by gp.word-to-the-wise.com (Postfix, from userid 500) id 8584190000D; Fri, 28 Nov 2003 12:37:00 -0800 (PST) Date: Fri, 28 Nov 2003 12:37:00 -0800 From: Steve Atkins To: pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or Message-ID: <20031128203700.GA19831@gp.word-to-the-wise.com> References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> <3FC54A82.2040605@familyhealth.com.au> <20031127032831.GA4836@gp.word-to-the-wise.com> <3FC58097.8090803@familyhealth.com.au> <20031127051230.GA6212@gp.word-to-the-wise.com> <20031128050417.GA14227@gp.word-to-the-wise.com> <3FC6DAB8.4080106@familyhealth.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3FC6DAB8.4080106@familyhealth.com.au> User-Agent: Mutt/1.4.1i X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/299 X-Sequence-Number: 4859 On Fri, Nov 28, 2003 at 01:18:48PM +0800, Christopher Kings-Lynne wrote: > > >Any thoughts? > > Actually, I ran my tests using tsearch V1. I wonder if there has been > some weird regression between tsearch 1 and 2? Maybe. tsearch2 doesn't seem production ready in other respects (untsearch2.sql barfs with 'aggregate stat(tsvector) does not exist' and the openfts mailing list, where this would be more appropriate, doesn't appear to exist according to sourceforge). So, using the same data, modulo a few alter tables, I try tsearch, V1. It's a little slower than V2, and again runs far faster without an index than with it. Broken in the same way. I have 7.2.4 running on a Sun box, so I tried that too, with similar results. tsearch just doesn't seem to work very well on this dataset (or any other large dataset I've tried). Cheers, Steve From pgsql-performance-owner@postgresql.org Sun Nov 30 05:47:11 2003 X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 27965D1B46B for ; Sun, 30 Nov 2003 09:47:08 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 20858-09 for ; Sun, 30 Nov 2003 05:46:37 -0400 (AST) Received: from relay01.kbs.net.au (relay01.kbs.net.au [203.220.32.149]) by svr1.postgresql.org (Postfix) with ESMTP id C9DEDD1B457 for ; Sun, 30 Nov 2003 05:46:35 -0400 (AST) Received: from [203.221.247.192] (helo=familyhealth.com.au) by relay01.kbs.net.au with esmtp (Exim 3.36 #1) id 1AQOA2-0005U9-00; Sun, 30 Nov 2003 20:46:38 +1100 Message-ID: <3FC9BC76.2030308@familyhealth.com.au> Date: Sun, 30 Nov 2003 17:46:30 +0800 From: Christopher Kings-Lynne User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031013 Thunderbird/0.3 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Steve Atkins Cc: pgsql-performance@postgresql.org Subject: Re: For full text indexing, which is better, tsearch2 or References: <4c0ccf4c27ff.4c27ff4c0ccf@jhmimail.jhmi.edu> <3FC54A82.2040605@familyhealth.com.au> <20031127032831.GA4836@gp.word-to-the-wise.com> <3FC58097.8090803@familyhealth.com.au> <20031127051230.GA6212@gp.word-to-the-wise.com> <20031128050417.GA14227@gp.word-to-the-wise.com> <3FC6DAB8.4080106@familyhealth.com.au> <20031128203700.GA19831@gp.word-to-the-wise.com> In-Reply-To: <20031128203700.GA19831@gp.word-to-the-wise.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200311/300 X-Sequence-Number: 4860 > I have 7.2.4 running on a Sun box, so I tried that too, with similar > results. tsearch just doesn't seem to work very well on this dataset > (or any other large dataset I've tried). Well, as I've shown - works fine for me... I strongly suggest you repost your problem report to -hackers, since the fact that the tsearch developers haven't chimed in implies to me that they don't watch the performance list. BTW, read this about Gist indexes: http://www.postgresql.org/docs/current/static/limitations.html (Note lack of concurrency) Chris