status
stringclasses 1
value | repo_name
stringlengths 9
24
| repo_url
stringlengths 28
43
| issue_id
int64 1
104k
| updated_files
stringlengths 8
1.76k
| title
stringlengths 4
369
| body
stringlengths 0
254k
⌀ | issue_url
stringlengths 37
56
| pull_url
stringlengths 37
54
| before_fix_sha
stringlengths 40
40
| after_fix_sha
stringlengths 40
40
| report_datetime
timestamp[ns, tz=UTC] | language
stringclasses 5
values | commit_datetime
timestamp[us, tz=UTC] |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
closed | onnx/onnx | https://github.com/onnx/onnx | 3,742 | ["onnx/common/ir.h"] | memory leaks issue | # Bug Report
exist memory leak by ASAN check
if I use this patch the issue will disappear
```
--- a/onnx/common/ir.h
+++ b/onnx/common/ir.h
@@ -1212,6 +1212,7 @@ private:
void freeValue(Value * v) {
auto it = all_values.find(v);
ONNX_ASSERT(it != all_values.end());
+ delete *it;
all_values.erase(it);
}
};
```
please anyone look at this issue, thank you
```
=================================================================
==27762==ERROR: LeakSanitizer: detected memory leaks
Direct leak of 9856 byte(s) in 77 object(s) allocated from:
#0 0x7f1fe341b448 in operator new(unsigned long) (/usr/lib/x86_64-linux-gnu/libasan.so.4+0xe0448)
#1 0x7f1fe303e8ab in onnx::Node::addOutput() (/home/chijinchao/workspace/GraphGen/build/core/import/libgraphgen_import.so+0x15e8ab)
#2 0x7f1fe3040d7d in onnx::Graph::create(onnx::Symbol, unsigned long) (/home/chijinchao/workspace/GraphGen/build/core/import/libgraphgen_import.so+0x160d7d)
#3 0x7f1fdf06213f in onnx::graphProtoToGraph(onnx::GraphProto const&, bool) third_party/onnx/onnx-src/onnx/common/ir_pb_converter.cc:243
#4 0x7f1fdf064210 in onnx::ImportModelProto(onnx::ModelProto const&) third_party/onnx/onnx-src/onnx/common/ir_pb_converter.cc:338
.......
SUMMARY: AddressSanitizer: 54784 byte(s) leaked in 588 allocation(s).
```
| https://github.com/onnx/onnx/issues/3742 | https://github.com/onnx/onnx/pull/3760 | a56a110c2c4e3b85114cb918b9c27dac211ab969 | 1897d1bed5a2cd56fcdfef3fd4e0ba2c4fe6a362 | 2021-09-28T08:36:32Z | python | 2021-10-19T05:50:33Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,740 | ["docs/Operators.md", "docs/TestCoverage.md", "onnx/backend/test/case/node/if.py", "onnx/backend/test/case/node/loop.py", "onnx/backend/test/data/node/test_if_opt/model.onnx", "onnx/backend/test/data/node/test_loop16_seq_none/model.onnx", "onnx/backend/test/data/node/test_loop16_seq_none/test_data_set_0/input_2.pb", "onnx/backend/test/data/node/test_loop16_seq_none/test_data_set_0/output_0.pb", "onnx/defs/shape_inference.cc", "onnx/defs/shape_inference.h"] | Loop operator's type inference doesn't handle propagating optional types from subgraph output to Loop output | # Bug Report
### Is the issue related to model conversion?
No
### Describe the bug
Please describe the bug clearly and concisely
When this line of code is hit for the optional type:
https://github.com/onnx/onnx/blob/9f978e1ffbd4ba29263f9fdae592e2256b80ec2a/onnx/defs/controlflow/defs.cc#L358
It breaks because the function doesn't support optional types yet:
https://github.com/onnx/onnx/blob/9f978e1ffbd4ba29263f9fdae592e2256b80ec2a/onnx/defs/shape_inference.h#L287
### System information
N/A
### Reproduction instructions
N/A
### Expected behavior
A clear and concise description of what you expected to happen.
### Notes
Any additional information
| https://github.com/onnx/onnx/issues/3740 | https://github.com/onnx/onnx/pull/3756 | 1259d16ff6715e78f3b7c0602d5dcb982b893383 | be76ca7148396176784ba8733133b9fb1186ea0d | 2021-09-27T20:04:42Z | python | 2021-10-20T18:50:15Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,565 | ["onnx/defs/tensor/old.cc"] | infer_shapes fails but onnxruntime works | # Bug Report
### Is the issue related to model conversion?
onnx raises an exception while running ``infer_shapes`` (``onnx.onnx_cpp2py_export.shape_inference.InferenceError: [ShapeInferenceError] (op_type:Sqrt, node name: ComplexAbsoutput__19): [ShapeInferenceError] Inferred shape and existing shape differ in dimension 1: (4) vs (3)``).
### Describe the bug
The following code fails (onnx graph: [bug_shape_inference.zip](https://github.com/onnx/onnx/files/6785879/bug_shape_inference.zip)):
```
import onnx
with open('bug_shape_inference.onnx', 'rb') as f: #see attached file
model_proto = onnx.load(f)
onnx.checker.check_model(model_proto, full_check=True)
```
Exception:
```
onnx.onnx_cpp2py_export.shape_inference.InferenceError: [ShapeInferenceError] (op_type:Sqrt, node name: ComplexAbsoutput__19): [ShapeInferenceError] Inferred shape and existing shape differ in dimension 1: (4) vs (3)
```
However, onnxruntime is able to compute the output:
```
from onnxruntime import InferenceSession
import numpy
mat = numpy.random.randn(3, 4).astype(numpy.float32)
sess = InferenceSession('bug_shape_inference.onnx')
print(sess.run(None, {'input:0': mat}))
```
It produces a matrix (3,3) as mentioned in the onnx graph:
```
output {
name: "output:0"
type {
tensor_type {
elem_type: 1
shape {
dim {
dim_value: 3
}
dim {
dim_value: 3
}
}
}
}
}
```
The error message seems to say onnx expects an output with shape (4,3) or (3,4) when it should be (3,3).
### System information
- OS Platform and Distribution (*Windows, Ubuntu 20.04*):
- ONNX version (*e.g. 1.7*): 1.9.0
- Python version: 3.9.5
- Visual Studio version (if applicable): 2019
### Reproduction instructions
See the bug description.
### Expected behavior
I would expect infer_shapes to succeed and find a shape equal to (3,3).
| https://github.com/onnx/onnx/issues/3565 | https://github.com/onnx/onnx/pull/3810 | 94fb49259018fc9ed95732bd7eb609bd518b0ee8 | f98b19a9390943d74ce7de30de1219aea8fff914 | 2021-07-08T16:19:38Z | python | 2021-12-08T21:29:24Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,562 | ["onnx/backend/test/runner/__init__.py"] | onnx-tensorflow onnx backend test of case test_identity_sequence_cpu failed | # Bug Report
### Is the issue related to model conversion?
No
### Describe the bug
When run onnx backend test case of test_identity_sequence_cpu with onnx-tensorflow, hit below error:
======================================================================
ERROR: test_identity_sequence_cpu (__main__.OnnxBackendNodeModelTest)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/opt/anaconda3/envs/otf/lib/python3.7/site-packages/onnx-1.9.0-py3.7-linux-x86_64.egg/onnx/backend/test/runner/__init__.py", line 250, in device_test_func
return test_func(*args, device=device, **kwargs)
File "/opt/anaconda3/envs/otf/lib/python3.7/site-packages/onnx-1.9.0-py3.7-linux-x86_64.egg/onnx/backend/test/runner/__init__.py", line 308, in run
atol=model_test.atol)
File "/opt/anaconda3/envs/otf/lib/python3.7/site-packages/onnx-1.9.0-py3.7-linux-x86_64.egg/onnx/backend/test/runner/__init__.py", line 172, in assert_similar_outputs
np.testing.assert_equal(outputs[i].dtype, ref_outputs[i].dtype)
AttributeError: 'list' object has no attribute 'dtype'
----------------------------------------------------------------------
Ran 1 test in 0.970s
FAILED (errors=1)
### System information
- OS Platform and Distribution (*e.g. Linux Ubuntu 16.04*):
- ONNX version 1.9.0 and master
- Python version:3.7
### Reproduction instructions
- Describe the code to reproduce the behavior.
```
import onnx
from onnx_tf.backend import TensorflowBackend
backend_test = onnx.backend.test.runner.Runner(TensorflowBackend, __name__)
if __name__ == '__main__':
unittest.main()
...
```
- Attach the ONNX model to the issue (where applicable)
### Expected behavior
The test case should pass (when onnx_tf implemented the operator correctly)
### Notes
The root cause is function assert_similar_outputs() in onnx/backend/test/runner/__init__.py does not handle sequence/list as model/node's output, it always check dtype of the output.
```
np.testing.assert_equal(outputs[i].dtype, ref_outputs[i].dtype)
```
| https://github.com/onnx/onnx/issues/3562 | https://github.com/onnx/onnx/pull/3563 | 21d16d2bcb493687eec2c2beefc7dde89d54a1e2 | cc7b5e7a5b7b569ec47f9ca36d379e1ea0b83c3b | 2021-07-08T10:00:52Z | python | 2021-07-12T20:50:35Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,530 | ["CMakeLists.txt"] | Multiple libraries (such as onnx) set WX when building with MSVC even when ONNX_WERROR is set to OFF | I understand from https://github.com/onnx/onnx/issues/1936 that this may be "by design", however, I see two problems with this:
1) It is inconsistent across platform AND a cursory look at the options leads one to believe that by setting ONNX_WERROR to OFF, warnings won't be treated as errors. This results (at least in my case) in a tracking down of where the issue that came from and time lost. At the very least, ONNX_WERROR should be documented correctly to indicate that even when it is OFF, WX will be passed to the compiler for some libraries.
2) Warnings are treated as errors that the user doesn't really have control over. For example, onnx-mlir includes onnx and uses the same version of protobuf as onnx or newer. When building onnx-mlir, we get the following warnings treated as errors (among others):
```
third_party\onnx\onnx\onnx-data.pb.cc(104): error C2220: the following warning is treated as an error
third_party\onnx\onnx\onnx-data.pb.cc(104): warning C4125: decimal digit terminates octal escape sequence
```
Since this is building unmodified onnx (1.9.0), this should be green, but it is not. So as the user of onnx, onnx-mlir has to explicitly disable those warnings or have build failures with MSVC.
| https://github.com/onnx/onnx/issues/3530 | https://github.com/onnx/onnx/pull/3548 | c79ab4c78599eb7f09cee8bf2d53af8250ed7dc9 | f24c062126225cddb49aecdedfebe7135b35ef50 | 2021-06-15T19:43:03Z | python | 2021-07-01T23:02:41Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,515 | ["onnx/defs/tensor/defs.cc"] | Squeeze shape inference failure with symbolic dimension | ```
import onnx
from onnx import version_converter, helper, TensorProto, shape_inference
import numpy as np
nodes = [helper.make_node('Squeeze', ['in', 'axes'], ['out'])]
inputs = [
helper.make_tensor_value_info('in', TensorProto.FLOAT, [1, 2, 'hello']),
helper.make_tensor_value_info('axes', TensorProto.INT64, [1])
]
outputs = [helper.make_tensor_value_info('out', TensorProto.FLOAT, (2, 'hello'))]
initializers = [helper.make_tensor('axes', TensorProto.INT64, [1], [0])]
model = helper.make_model(helper.make_graph(nodes, "g", inputs, outputs, initializers))
shape_inference.infer_shapes(model)
```
Shape inference fails with
```
(op_type:Squeeze): [ShapeInferenceError] Inferred shape and existing shape differ in rank: (0) vs (2)
```
Seems like this stopped working with https://github.com/onnx/onnx/pull/3465 | https://github.com/onnx/onnx/issues/3515 | https://github.com/onnx/onnx/pull/3516 | d08b3e951be607b9638ab84c340fabec1fdbec83 | 7a2bd3567762c0bc6b941c70fa43ba780a192f27 | 2021-06-04T16:40:25Z | python | 2021-06-15T18:42:47Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,505 | ["onnx/defs/quantization/defs.cc", "onnx/test/shape_inference_test.py"] | DynamicQuantizeLinear function does not have shape inference function | # Bug Report
### Is the issue related to model conversion? No
### Describe the bug
DynamicQuantizeLinear function op does not have shape inference function defined. In absence of shape inference, function body is used to get the shape inference for the function op and although it works as a fallback option it hurts perf.
### Expected behavior
Add shape inference function for DynamicQuantizeLinear
### Notes
Any additional information
| https://github.com/onnx/onnx/issues/3505 | https://github.com/onnx/onnx/pull/3539 | e6ea9356fb4ba62515fa1cc9591eab5130c2e5ab | 3c857c57826cf434626cb8c99f0b9517efb62451 | 2021-05-27T06:25:16Z | python | 2021-06-22T17:27:51Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,439 | ["onnx/defs/tensor/defs.cc", "onnx/shape_inference/implementation.cc"] | Shape inference for Unique assigns shape to empty string when optional outputs are provided. | A value_info entry is added for empty string, which is not correct.
Onnx version 1.8.1.
https://www.dropbox.com/s/ljyo821c0b2miv5/test_unsorted_segment_sum.onnx?dl=1 | https://github.com/onnx/onnx/issues/3439 | https://github.com/onnx/onnx/pull/3815 | 34dbc3e3a8f4709d96e7ceba3ac38510f4ed2ff3 | fa6f8cfdce3d86346e8a7494f3062b98416c85fb | 2021-04-19T23:21:26Z | python | 2021-11-22T22:40:13Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,428 | ["docs/Operators.md", "docs/TestCoverage.md", "onnx/backend/test/case/node/roialign.py", "onnx/backend/test/data/node/test_roialign/test_data_set_0/output_0.pb", "onnx/backend/test/data/node/test_roialign_identity_region/model.onnx", "onnx/backend/test/data/node/test_roialign_identity_region/test_data_set_0/input_0.pb", "onnx/backend/test/data/node/test_roialign_identity_region/test_data_set_0/input_1.pb", "onnx/backend/test/data/node/test_roialign_identity_region/test_data_set_0/input_2.pb", "onnx/backend/test/data/node/test_roialign_identity_region/test_data_set_0/output_0.pb"] | [Operator] RoiAlign backend test case's output is misaligned | # Bug Report
### Is the issue related to model conversion?
No.
### Describe the bug
The [RoiAlign operator](https://github.com/onnx/onnx/blob/master/docs/Operators.md#RoiAlign), per the [Mask R-CNN paper](https://arxiv.org/abs/1703.06870) and Facebook Research's [Detectron 2 implementation](https://github.com/facebookresearch/detectron2/blob/master/detectron2/layers/roi_align.py) aligns sampling points over the center of the pixels, but the [solitary test case](https://github.com/onnx/onnx/blob/master/onnx/backend/test/case/node/roialign.py) in the ONNX backend has output that is misaligned by half an input pixel. After comparing to various references (table below under **Notes**) and reading more into the history of this operator, I see this is actually an old bug (which Detectron since fixed and PyTorch also fixed with aligned=True) which applied an offset the output subsample by 0.5 but forgot to adjust the input sample to compensate (see their comment in the [code](https://github.com/facebookresearch/detectron2/blob/master/detectron2/layers/roi_align.py): "_the original roi_align (aligned=False) does not subtract the 0.5 when computing neighboring pixel indices and therefore it uses pixels with a slightly incorrect alignment (relative to our pixel model) when performing bilinear interpolation_" and the [test case](https://github.com/facebookresearch/detectron2/blob/master/tests/layers/test_roi_align.py#L15)).
| equation
----------|----------
Correct | `input.x = (output.x + 0.5) * computedOutputToInputScale + scaledCurrentRoi.x - 0.5` |
Incorrect | `input.x = (output.x + 0.5) * computedOutputToInputScale + scaledCurrentRoi.x - 0.0` |
We should fix the existing [ONNX backend test case](https://github.com/onnx/onnx/blob/master/onnx/backend/test/case/node/roialign.py) accordingly. Otherwise even a simple case of cropping an identity region from the input doesn't match. e.g. a new test case:
```python
# Crop a 2x2 region from the 4x5 input.
@staticmethod
def export_roialign_identity_region(): # type: () -> None
node = onnx.helper.make_node(
"RoiAlign",
inputs=["X", "rois", "batch_indices"],
outputs=["Y"],
spatial_scale=1.0,
output_height=2,
output_width=2,
sampling_ratio=1,
)
X = np.array(
[
[
[
[ 0, 1, 2, 3],
[ 4, 5, 6, 7],
[ 8, 9,10,11], # 2x2 region of interest: 9,10
[12,13,14,15], # 13,14
[16,17,18,19],
]
]
],
dtype=np.float32,
)
batch_indices = np.array([0], dtype=np.int64)
rois = np.array([[1, 2, 3, 4]], dtype=np.float32)
# (num_rois, C, output_height, output_width)
Y = np.array(
[
[
[
[ 9,10], # Should be identity scale/translation window into the input,
[13,14], # and not [[11.5,12.5], [15.5,16.5]].
]
]
],
dtype=np.float32,
)
expect(node, inputs=[X, rois, batch_indices], outputs=[Y], name="test_roialign_identity_region")
```
[roi_align_identity_region.zip](https://github.com/onnx/onnx/files/6322528/roi_align_identity_region.zip)
### System information
- OS Platform and Distribution: Windows 10 20H2 19042.867
- ONNX version: 1.8.1
- Python version: 3.9.1
- GCC/Compiler version: VS2019
- CMake version: NA
- Protobuf version: NA
- Visual Studio version (if applicable): VS2019
### Reproduction instructions
- Run the existing ONNX back test case.
### Expected behavior
The test case should bilinearly sample from pixel centers in *both* input and output pixels, not half of one and zero of the other.
e.g.
### Notes
I compared the bounding box output for the [Mask R-CNN model](https://github.com/onnx/models/tree/master/vision/object_detection_segmentation/mask-rcnn/model) using the baseline output in the tar.gz vs the half-pixel aligned output from the ONNX Runtime DirectML execution provider, and I feel the alignment is tighter and more accurately aligned, with fewer random false positives in the half-pixel sampled implementation. Other data scientists have also complained online that the old bug in PyTorch negatively affected their object detection results.
### Framework comparisons
```
Input Tensor =
<------->
0.0 1.0 2.0 3.0 4.0 5.0 6.0
|.5 |.5 |.5 |.5 |.5 |.5 |
0.0___ |_|_|_|_|_|_|_|_|_|_|_|_|
1.0___[| 0,| 1,| 2,| 3,| 4,| 5 ]
/|\ 2.0___[|10,┃11,┃12,┃13,|14,|15 ]
\|/ 3.0___[|20,┃21,┃22,┃23,|24,|25 ]
4.0___[|30,|31,|32,|33,|34,|35 ]
5.0___[|40,|41,|42,|43,|44,|45 ]
6.0___[|50,|51,|52,|53,|54,|55 ]
Active region of interest = [[1.0, 1.0, 3.0, 3.0], ...]
// The first window is 2x2 over the input elements, covering values [[11,12],[21,22]].
Output tensor size = [4,4]
```
Image | Source | Output 4x4, from first 2x2 region
-- | -- | --
 | ✔ FB Research [Detectron 2](https://github.com/facebookresearch/detectron2/blob/master/detectron2/layers/roi_align.py) (MaskedRCNN paper) | [ 8.25, 8.75, 9.25, 9.75],<br/>[13.25, 13.75, 14.25, 14.75],<br/>[18.25, 18.75, 19.25, 19.75],<br/>[23.25, 23.75, 24.25, 24.75]
 | ✔ ONNX Runtime DirectML EP (ROI_ALIGN 0) | [ 8.25, 8.75, 9.25, 9.75],<br/>[13.25, 13.75, 14.25, 14.75],<br/>[18.25, 18.75, 19.25, 19.75],<br/>[23.25, 23.75, 24.25, 24.75]
 | ✔ ONNX Runtime 1.7 CPU [Resize](https://github.com/onnx/onnx/blob/master/docs/Operators.md#Resize) + Slice<br/>coordinate_transformation_mode=half | [ 8.25, 8.75, 9.25, 9.75],<br/>[13.25, 13.75, 14.25, 14.75],<br/>[18.25, 18.75, 19.25, 19.75],<br/>[23.25, 23.75, 24.25, 24.75]
 | ✔ [`torchvision.ops.roi_align(aligned=True…)`](https://pytorch.org/vision/0.8/_modules/torchvision/ops/roi_align.html) | [ 8.25, 8.75, 9.25, 9.75],<br/>[13.25, 13.75, 14.25, 14.75],<br/>[18.25, 18.75, 19.25, 19.75],<br/>[23.25, 23.75, 24.25, 24.75]
 | ✖ [`torchvision.ops.roi_align(aligned=False…)`](https://pytorch.org/vision/0.8/_modules/torchvision/ops/roi_align.html)<br/>*deprecated, legacy flag still exists | [13.75, 14.25, 14.75, 15.25],<br/>[18.75, 19.25, 19.75, 20.25],<br/>[23.75, 24.25, 24.75, 25.25],<br/>[28.75, 29.25, 29.75, 30.25]
 | ✖ ONNX Runtime 1.7 CPU EP RoiAlign<br/>[PR \#7354](https://github.com/microsoft/onnxruntime/pull/7354) [Issue \#6921](https://github.com/microsoft/onnxruntime/issues/6921) | [13.75, 14.25, 14.75, 15.25],<br/>[18.75, 19.25, 19.75, 20.25],<br/>[23.75, 24.25, 24.75, 25.25],<br/>[28.75, 29.25, 29.75, 30.25]
 | 🤨 [`tf.image.crop_and_resize(…)`](https://www.tensorflow.org/api_docs/python/tf/image/crop_and_resize)<br/>*Note boxes are normalized 0 to 1 (so /5 each ROI element) | [11.00, 11.66, 12.33, 13.00],<br/>[17.66, 18.33, 19.00, 19.66],<br/>[24.33, 25.00, 25.66, 26.33],<br/>[31.00, 31.66, 32.33, 33.00]
 | 🤨 `tf.image.resize_bilinear(align_corners=True…)`<br/> + `tf.slice` | [11.00, 11.66, 12.33, 13.00],<br/>[17.66, 18.33, 19.00, 19.66],<br/>[24.33, 25.00, 25.66, 26.33],<br/>[31.00, 31.66, 32.33, 33.00]
 | 🤨 `tf.image.resize_bilinear(align_corners=False…)`<br/> + `tf.slice` | [11.00, 11.50, 12.00, 12.50],<br/>[16.00, 16.50, 17.00, 17.50],<br/>[21.00, 21.50, 22.00, 22.50],<br/>[26.00, 26.50, 27.00, 27.50]
 | ✔ `tf.image.resize_bilinear(half_pixel_centers=True…)`<br/> + `tf.slice` | [ 8.25, 8.75, 9.25, 9.75],<br/>[13.25, 13.75, 14.25, 14.75],<br/>[18.25, 18.75, 19.25, 19.75],<br/>[23.25, 23.75, 24.25, 24.75]
----------
Facebook research's [Detectron 2 test code](https://github.com/facebookresearch/detectron2/blob/master/tests/layers/test_roi_align.py#L15):
```python
class ROIAlignTest(unittest.TestCase):
def test_forward_output(self):
input = np.arange(25).reshape(5, 5).astype("float32")
"""
0 1 2 3 4
5 6 7 8 9
10 11 12 13 14
15 16 17 18 19
20 21 22 23 24
"""
output = self._simple_roialign(input, [1, 1, 3, 3], (4, 4), aligned=False)
output_correct = self._simple_roialign(input, [1, 1, 3, 3], (4, 4), aligned=True)
# without correction:
old_results = [
[7.5, 8, 8.5, 9],
[10, 10.5, 11, 11.5],
[12.5, 13, 13.5, 14],
[15, 15.5, 16, 16.5],
]
# with 0.5 correction:
correct_results = [
[4.5, 5.0, 5.5, 6.0],
[7.0, 7.5, 8.0, 8.5],
[9.5, 10.0, 10.5, 11.0],
[12.0, 12.5, 13.0, 13.5],
]
# This is an upsampled version of [[6, 7], [11, 12]]
...
```
----------
PyTorch sample code:
```python
# pip install torch==1.7.1+cpu torchvision==0.8.2+cpu torchaudio===0.7.2 -f https://download.pytorch.org/whl/torch_stable.html
import torch
import torchvision
print("PyTorch version:", torch.__version__)
input = [[[[ 0, 1, 2, 3, 4, 5], # NCHW
[10,11,12,13,14,15],
[20,21,22,23,24,25],
[30,31,32,33,34,35],
[40,41,42,43,44,45],
[50,51,52,53,54,55]]]]
boxes = [[0, 1,1,3,3]]
output_size = [4,4]
aligned=True # Correct
#aligned=False # Legacy setting
sampling_ratio=1
spatial_scale=1
# https://pytorch.org/vision/0.8/_modules/torchvision/ops/roi_align.html
output = torchvision.ops.roi_align(
torch.tensor(input, dtype=torch.float),
torch.tensor(boxes, dtype=torch.float),
output_size,
spatial_scale=spatial_scale,
sampling_ratio=sampling_ratio,
aligned=aligned
)
torch.set_printoptions(sci_mode=False)
print(input)
print(boxes)
print(output)
```
----------
TensorFlow sample code:
```python
# pip install tensorflow-gpu==1.15.0
import os
import tensorflow.compat.v1 as tf
input = [[ # NHWC
[[ 0.], [ 1.], [ 2.], [ 3.], [ 4.], [ 5.]],
[[10.], [11.], [12.], [13.], [14.], [15.]],
[[20.], [21.], [22.], [23.], [24.], [25.]],
[[30.], [31.], [32.], [33.], [34.], [35.]],
[[40.], [41.], [42.], [43.], [44.], [45.]],
[[50.], [51.], [52.], [53.], [54.], [55.]]
]]
boxes = [[1/5,1/5,3/5,3/5],[3/5,3/5,4/5,4/5]] # Normalized 0.0 to 1.0 (where 1.0 = width - 1 and height - 1)
box_indices = [0, 0] # Batch indices per corresponding region
crop_size = [4, 4] # Output tensor size HW
print("TensorFlow version:", tf.__version__) # 1.15.0 (cpu/cuda)
# Using half_pixel_centers=True is correct (not align_corners=True)
output_size = [6*2, 6*2]
resize_output = tf.image.resize_bilinear(tf.constant(input), output_size, align_corners=False, half_pixel_centers=True)
resize_bilinear_slice_output = tf.slice(resize_output, [0,2,2,0], [1,4,4,1])
# Note crop_and_resize doesn't scale the image boundaries to pixel centers, but always to corners,
# and there is sadly no flag to influence this (unlike resize_bilinear).
method = 'bilinear'
extrapolation_value = 0
crop_and_resize_output = tf.image.crop_and_resize(
image=tf.constant(input, dtype=tf.float32), # NHWC
boxes=tf.constant(boxes, dtype=tf.float32),
box_ind=tf.constant(box_indices, dtype=tf.int32),
crop_size=tf.constant(crop_size, dtype=tf.int32),
method=method,
extrapolation_value=extrapolation_value
)
with tf.Session(config=config) as session:
with np.printoptions(precision=3, suppress=True):
print("input:\n", input)
print("crop_and_resize:\n", session.run(crop_and_resize_output))
print("resize_bilinear_and_slice:\n", session.run(resize_bilinear_slice_output))
```
| https://github.com/onnx/onnx/issues/3428 | https://github.com/onnx/onnx/pull/3429 | 1880e91cd899076240117efb594fa4c9ab144f2c | 82ddc566b950d05fdccb70c4579930f4014ece7f | 2021-04-16T05:45:17Z | python | 2021-04-30T22:06:56Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,409 | ["docs/Changelog.md", "docs/Operators.md", "onnx/defs/math/defs.cc", "onnx/defs/math/old.cc", "onnx/defs/operator_sets.h", "onnx/defs/schema.h"] | Pow support for bfloat16 | I see that with opset13 bfloat16 support has been added to [Pow](https://github.com/onnx/onnx/blob/master/docs/Operators.md#pow
), however only the first operand `T` has been updated to support bfloat16 ; should the second operand `T1` be updated as well?
Reproduced here from https://github.com/onnx/onnx/blob/master/docs/Operators.md#pow
```
T : tensor(int32), tensor(int64), tensor(float16), tensor(float), tensor(double), tensor(bfloat16)
Constrain input X and output types to float/int tensors.
T1 : tensor(uint8), tensor(uint16), tensor(uint32), tensor(uint64), tensor(int8), tensor(int16), tensor(int32), tensor(int64), tensor(float16), tensor(float), tensor(double)
Constrain input Y types to float/int tensors.
```
| https://github.com/onnx/onnx/issues/3409 | https://github.com/onnx/onnx/pull/3412 | fe1d4946ccde54669dbabb47d4bec0aa14e92471 | a64c2ed232392e2586dc8320710841eee62874df | 2021-04-07T19:16:18Z | python | 2021-04-24T23:56:26Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,345 | [".azure-pipelines/Linux-CI.yml", ".azure-pipelines/Windows-CI.yml", ".github/workflows/manylinux/entrypoint.sh", ".github/workflows/release_win.yml", ".github/workflows/weekly_mac_ci.yml", ".github/workflows/win_no_exception_ci.yml", "CMakeLists.txt", "README.md", "cmake/summary.cmake"] | ONNX_USE_PROTOBUF_SHARED_LIBS should control Protobuf_USE_STATIC_LIBS | # Bug Report
### Describe the bug
ONNX_USE_PROTOBUF_SHARED_LIBS is set independently of Protobuf_USE_STATIC_LIBS even though the two are dependent on each other. Even the documentation says:
If ONNX_USE_PROTOBUF_SHARED_LIBS is ON then Protobuf_USE_STATIC_LIBS must be OFF and USE_MSVC_STATIC_RUNTIME must be 0. If ONNX_USE_PROTOBUF_SHARED_LIBS is OFF then Protobuf_USE_STATIC_LIBS must be ON.
Rather than relying on the user to set Protobuf_USE_STATIC_LIBS when calling cmake, it should be set before the call to find_package for protobuf based on the value of ONNX_USE_PROTOBUF_SHARED_LIBS. This way the user does not have to rely on their knowledge of what cmake does when find_package is called.
### System information
- OS Platform and Distribution (*e.g. Linux Ubuntu 16.04*): Windows 10
- ONNX version (*e.g. 1.7*): 1.7
- Python version: 3.8
- GCC/Compiler version (if compiling from source): VS2019
- CMake version: 3.19
- Protobuf version: 3.12.3
- Visual Studio version (if applicable): VS2019
### Reproduction instructions
1. Checkout protobuf 3.12.3
2. Correctly build static libraries with cmake and VS2019
-G "Visual Studio 16 2019" -A x64 -T host=x64
-DCMAKE_BUILD_TYPE=Release
-DCMAKE_INSTALL_PREFIX=/path/to/protobuf/install
-Dprotobuf_BUILD_EXAMPLES=OFF
-Dprotobuf_BUILD_TESTS=OFF
-Dprotobuf_MSVC_STATIC_RUNTIME=OFF
-Dprotobuf_WITH_ZLIB=OFF
3. Checkout onnx-mlir
4. Build with cmake and VS2019
-G "Visual Studio 16 2019" -A x64 -T host=x64
-DCMAKE_BUILD_TYPE=Release
-DCMAKE_PREFIX_PATH=/path/to/protobuf/install
### Expected behavior
Build succeeds
### Actual behavior
The build fails with multiple linker errors:
LNK2001: unresolved external symbol "__declspec(dllimport) ..."
LNK2019: unresolved external symbol "__declspec(dllimport) ..."
because neither onnx-mlir, nor onnx define Protobuf_USE_STATIC_LIBS and cmake silently defines PROTOBUF_USE_DLLS, but our protobuf is a static library.
| https://github.com/onnx/onnx/issues/3345 | https://github.com/onnx/onnx/pull/3550 | 6c77a411215305c1a3f355f8c2e4633c19514690 | 31f0fd6103e231664d1cfd3b25264f672f5de960 | 2021-03-20T00:37:41Z | python | 2021-07-02T21:24:44Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,328 | ["onnx/test/version_converter_test.py", "onnx/version_converter/convert.cc", "onnx/version_converter/convert.h"] | Version converter doesn't convert nodes in subgraphs | The following script compares updating an `Unsqueeze` operator from a top-level graph vs an `Unsqueeze` operator from the body attribute of a `Loop` operator. It uses files from the node tests (update value of `node_tests_dir` to reproduce):
```
import onnx
from onnx import version_converter, shape_inference
node_tests_dir = '/home/matteo/Git/onnx/onnx/backend/test/data/node/'
unsqueeze_test = 'test_unsqueeze_axis_3/'
loop_test = 'test_loop11/'
print('**** Unsqueeze model from Unsqueeze test')
path = node_tests_dir + unsqueeze_test + 'model.onnx'
model = onnx.load(path)
print('Before conversion:')
print('opset version: ' + str(model.opset_import[0].version))
print(model.graph.node[0])
print('After conversion:')
converted = version_converter.convert_version(model, 13)
print('opset version: ' + str(converted.opset_import[0].version))
print(converted.graph.node[0])
print('')
print('**** Unsqueeze model from Loop test')
path = node_tests_dir + loop_test + 'model.onnx'
model = onnx.load(path)
print('Before conversion:')
print('opset version: ' + str(model.opset_import[0].version))
print(model.graph.node[0].attribute[0].g.node[4])
print('After conversion:')
converted = version_converter.convert_version(model, 13)
print('opset version: ' + str(converted.opset_import[0].version))
print(converted.graph.node[0].attribute[0].g.node[4])
print('')
print('**** Shape inference on converted Loop model')
shape_inference.infer_shapes(converted)
```
The output is:
```
**** Unsqueeze model from Unsqueeze test
Before conversion:
opset version: 11
input: "x"
output: "y"
op_type: "Unsqueeze"
attribute {
name: "axes"
ints: 3
type: INTS
}
After conversion:
opset version: 13
input: "x"
input: "4"
output: "y"
op_type: "Unsqueeze"
**** Unsqueeze model from Loop test
Before conversion:
opset version: 11
input: "iter_count"
output: "slice_start"
op_type: "Unsqueeze"
attribute {
name: "axes"
ints: 0
type: INTS
}
After conversion:
opset version: 13
input: "iter_count"
output: "slice_start"
op_type: "Unsqueeze"
attribute {
name: "axes"
ints: 0
type: INTS
}
**** Shape inference on converted Loop model
Traceback (most recent call last):
File "script.py", line 33, in <module>
shape_inference.infer_shapes(converted)
File "/home/matteo/.local/lib/python3.8/site-packages/onnx-1.8.1-py3.8-linux-x86_64.egg/onnx/shape_inference.py", line 37, in infer_shapes
inferred_model_str = C.infer_shapes(model_str, check_type, strict_mode)
RuntimeError: input 1 is out of bounds
```
It can be seen that in the second case the `Unsqueeze` operator has not been updated, producing a faulty model which makes shape inference fail.
So it seems that the version converter doesn't bother to recurse into the `Loop` subgraphs. Is this a problem with every operator supporting subgraphs? | https://github.com/onnx/onnx/issues/3328 | https://github.com/onnx/onnx/pull/3474 | 9373cfc412b4d4ed54ea0e6d32aaf3722ee30493 | ada1eed304bebbb441f732d583f118825eefa0fa | 2021-03-14T15:40:09Z | python | 2021-05-18T17:19:31Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,314 | ["docs/Changelog.md", "docs/Operators.md", "docs/TestCoverage.md", "onnx/backend/test/case/node/averagepool.py", "onnx/backend/test/case/node/lppool.py", "onnx/backend/test/case/node/maxpool.py", "onnx/backend/test/case/node/pool_op_common.py", "onnx/defs/nn/defs.cc", "onnx/defs/nn/old.cc"] | Confused about AveragePool op ceil_mode attribute | # Ask a Question
### Question
When I was study the AveragePool op, I am confused about the ceil_mode attr.
What I see at [AveragePool](https://github.com/onnx/onnx/blob/master/docs/Operators.md#AveragePool)
> ceil_mode : int (default is 0)
> Whether to use ceil or floor (default) to compute the output shape.
But I notice when auto_pad is set,
it only uses ceil to compute the output shape.
Is the ceil_mode attr only used when auto_pad is 'NOTSET'?
### Further information
- This is the python code I find to compute the output shape, the code also is only use ceil to compute output shape. [get_output_shape](https://github.com/onnx/onnx/blob/1cbd549193661c7e9c870c4e9595144d0c6971c5/onnx/backend/test/case/node/pool_op_common.py#L24)
- I notice pytorch's op AvgPool2d have ceil_mode( [AvgPool2d,](https://pytorch.org/docs/stable/generated/torch.nn.AvgPool2d.html) )I guess the onnx ceil_mode attr is only used when auto_pad is 'NOTSET', and when auto_pad is not 'NOTSET',the ceil_mode default value is 1 ,is true?
Who can help me? Thanks very much.
| https://github.com/onnx/onnx/issues/3314 | https://github.com/onnx/onnx/pull/5248 | efa11e6461960825ae5fc23ce48893ac25853bad | 57057ac48f1b87c01789888e11a3c2bf8852e12c | 2021-03-05T06:15:49Z | python | 2023-06-15T04:28:58Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,255 | ["onnx/test/checker_test.py"] | Not able to run inference with If operator | # Bug Report
### Is the issue related to model conversion?
No
### Describe the bug
https://github.com/onnx/onnx/blob/master/onnx/test/checker_test.py#L254
We created a model based on the above test case (test_nested_graph). Onnx model was successfully created but when running inference based on the generated model we got the following error:
The code we're running:
`model_path = "models/test1.onnx"`
`input_data = [np.array([True]), np.array([1.0, 2.0]).astype(np.float32)]`
`onnx_sess = onnxruntime.InferenceSession(model_path)`
`pred = onnx_sess.run(None, input_data)`
`print(pred)`
Traceback (most recent call last):
File "create_test_model.py", line 299, in <module>
onnx_sess = onnxruntime.InferenceSession(model_out_path)
File "lib/python3.7/site-packages/onnxruntime/capi/session.py", line 195, in __init__
self._create_inference_session(providers, provider_options)
File "lib/python3.7/site-packages/onnxruntime/capi/session.py", line 200, in _create_inference_session
sess = C.InferenceSession(session_options, self._model_path, True, self._read_config_from_model)
onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Load model from models/test1.onnx failed:Node () Op (If) [TypeInferenceError] Graph attribute inferencing failed: Size mismatch validating subgraph inputs. Got 0 inputs but subgraph has 1 inputs and requires 1 inputs. Either provide all subgraph inputs, or just the required inputs.
### System information
- OS Platform and Distribution: Mac OS 10.14.6
- ONNX version: 1.7.0
- Python version: 3.7.8
- GCC/Compiler version (if compiling from source):
- CMake version:
- Protobuf version:
- Visual Studio version (if applicable):
### Reproduction instructions
- Describe the code to reproduce the behavior.
See description.
- Attach the ONNX model to the issue (where applicable)
### Expected behavior
Expect onnxsession to be created and inference session to return float result
### Notes
Any additional information
| https://github.com/onnx/onnx/issues/3255 | https://github.com/onnx/onnx/pull/3798 | d0151d78c2ba53c3c71d05608b1d892ada20507d | 5f1a8079561a0b677c9b8f82ecabff368051e79d | 2021-02-02T01:44:50Z | python | 2021-10-26T16:45:41Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,237 | ["onnx/backend/test/report/coverage.py"] | Operator coverage report broken in ONNX 1.8 | # Bug Report
### Describe the bug
I’m trying to debug a problem we noticed on the [ONNX Scoreboard](https://github.com/onnx/backend-scoreboard). For a while now the operator coverage report has not been updating. While debugging this I noticed that the coverage report is broken because it can’t find a schema for Adagrad (and perhaps other ops). The error message I’m seeing is this:
```
RuntimeError: No schema registered for 'Adagrad'!
```
You can see [complete logs on Azure](https://dev.azure.com/michalkarzynski/michalkarzynski/_build/results?buildId=995&view=logs&j=89a03e46-fbf8-5c6e-11d2-f1a7ba9bbbde&t=d35e1a0e-28e7-52d8-f96e-e1252e0b4d87&l=4748).
### System information
- OS Platform and Distribution: Linux Ubuntu 20.04
- ONNX version 1.8
- Python version: Python 3.6.9
### Reproduction instructions
You can download the [Dockerfile for ONNX Runtime](https://github.com/onnx/backend-scoreboard/blob/master/runtimes/onnx-runtime/stable/Dockerfile) from ONNX Scoreboard, or checkout the entire repo.
```
docker build -t scoreboard/onnx -f runtimes/onnx-runtime/stable/Dockerfile .
docker run --name onnx-runtime --env-file setup/env.list -v `pwd`/results/onnx-runtime/stable:/root/results scoreboard/onnx
```
Internally this runs:
```
pytest test_backend.py --onnx_backend=onnxruntime.backend.backend -k 'not OnnxBackendRealModelTest' --tb=no
```
Using the following [test file](https://github.com/onnx/backend-scoreboard/blob/master/test/test_backend.py).
### Expected behavior
We would expect to see a file named [`nodes.csv`](https://github.com/onnx/backend-scoreboard/blob/master/results/onnx-runtime/stable/nodes.csv) to be generated.
### Notes
Instead we see `RuntimeError: No schema registered for 'Adagrad'!`.
Why is Adagrad schema is not being registered for the operator coverage?
| https://github.com/onnx/onnx/issues/3237 | https://github.com/onnx/onnx/pull/3256 | 6544dba74a4a0c9d937507b6f8f6a4cb20234701 | b4d69b49f4216ebbe8857bd76e7c1a77665ee44f | 2021-01-25T16:17:51Z | python | 2021-02-08T09:42:34Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,228 | ["docs/ExternalData.md", "docs/PythonAPIOverview.md", "onnx/__init__.py", "onnx/external_data_helper.py", "onnx/test/test_external_data.py"] | Updates to external data helpers | Add a new API for converting a model to external data. Today the conversion happens in 2 steps
`external_data_helper.convert_model_to_external_data(<model>,
<all_tensors_to_one_file>, <size_threshold>)
save_model(model, output_path)
`
We want to add another api which combines the 2 steps
`
save_model_to_external_data(<model>, <output_path>, <all_tensors_to_one_file>, <size_threshold>, <convert_attributes>)
`
| https://github.com/onnx/onnx/issues/3228 | https://github.com/onnx/onnx/pull/3280 | 39df720cca9c4a833a3de3a6bc1d4667c1a8e58b | 522b000546e5ed986fb2724ab2099d5a3219c586 | 2021-01-19T23:18:14Z | python | 2021-03-24T20:02:31Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,186 | ["onnx/defs/shape_inference.cc", "onnx/defs/shape_inference.h", "onnx/shape_inference/implementation.cc", "onnx/shape_inference/implementation.h"] | Opset conversion | # Bug Report
### Is the issue related to model conversion?
Yes and no. Yes, in the sense that I've converted a model to ONNX. No, because I'm now trying to convert the opset to 7
### Describe the bug
I've converted a multiclass model to onnx with the target opset set to 7. When I tried to use it as with asp.net, I got an error saying it was opset 1 and that I needed to convert it. I'm now trying to convert it from opset 1 to opset 7 but I'm getting
```
RuntimeError: [TypeInferenceError] type case unsupported. existing=5 inferred=5
(op_type:ZipMap, node name: ZipMap): [TypeInferenceError] type case unsupported. existing=5 inferred=5
```
**Code used to convert sklearn model**
```
initial_type = [('float_input', FloatTensorType([1, dim]))]
onnx_model = convert_sklearn(lr, initial_types=initial_type, target_opset=12)
with open("Modele/LR_2.onnx", "wb") as f:
f.write(onnx_model.SerializeToString())
```
**Code used to convert opset**
```
original_model = onnx.load(model_path)
print('The model before conversion:\n{}'.format(original_model))
# Apply the version conversion on the original model
converted_model = version_converter.convert_version(original_model, 7)
print('The model after conversion:\n{}'.format(converted_model))
```
**Output :**
```
floats: 0.9884291
floats: 1.063458
floats: 0.6337186
floats: -0.42541054
floats: -0.0067295996
floats: 0.34725153
floats: 0.0008511626
floats: 0.009505493
floats: 0.18251349
floats: -0.1303715
type: FLOATS
}
attribute {
name: "multi_class"
i: 1
type: INT
}
attribute {
name: "post_transform"
s: "SOFTMAX"
type: STRING
}
domain: "ai.onnx.ml"
}
node {
input: "probability_tensor"
output: "probabilities"
name: "Normalizer"
op_type: "Normalizer"
attribute {
name: "norm"
s: "L1"
type: STRING
}
domain: "ai.onnx.ml"
}
node {
input: "label"
output: "output_label"
name: "Identity"
op_type: "Identity"
domain: ""
}
node {
input: "probabilities"
output: "output_probability"
name: "ZipMap"
op_type: "ZipMap"
attribute {
......
type: STRINGS
}
domain: "ai.onnx.ml"
}
name: "c50b4cd9856940b3924f51048cff27c5"
input {
name: "float_input"
type {
tensor_type {
elem_type: 1
shape {
dim {
dim_value: 1
}
dim {
dim_value: 14525
}
}
}
}
}
output {
name: "output_label"
type {
tensor_type {
elem_type: 8
shape {
dim {
}
}
}
}
}
output {
name: "output_probability"
type {
sequence_type {
elem_type {
map_type {
key_type: 8
value_type {
tensor_type {
elem_type: 1
}
}
}
}
}
}
}
}
opset_import {
domain: ""
version: 1
}
opset_import {
domain: "ai.onnx.ml"
version: 1
}
Traceback (most recent call last):
converted_model = version_converter.convert_version(original_model, 7)
converted_model_str = C.convert_version(model_str, target_version)
RuntimeError: [TypeInferenceError] type case unsupported. existing=5 inferred=5
(op_type:ZipMap, node name: ZipMap): [TypeInferenceError] type case unsupported. existing=5 inferred=5
``` | https://github.com/onnx/onnx/issues/3186 | https://github.com/onnx/onnx/pull/3772 | 780b03610021d58c166b3c113b849899845421db | f9a21106141940ec89fb338895cbc4f573344e0d | 2020-12-28T11:46:18Z | python | 2021-10-28T04:28:26Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,084 | ["CMakeLists.txt", "third_party/pybind11"] | Cannot build with pybind11 2.6.0 | # Bug Report
### Describe the bug
We cannot build with pybind11 2.6 released recently, 2.5 works fine.
Basically a `-I` for `Python.h` is missing.
A bug is filed to them: https://github.com/pybind/pybind11/issues/2632
They mentioned it could be caused by the wrong use in onnx's cmake.
Could you take a look?
### System information
- OS Platform and Distribution (*e.g. Linux Ubuntu 16.04*): CentOS 7 / Ubuntu 18.04
- ONNX version (*e.g. 1.7*): Built from latest release tag (1.7.0)
- Python version: distro stock python3
- GCC/Compiler version (if compiling from source): gcc-9 / gcc-8
- CMake version: Built from latest release tag
- Protobuf version: Built from latest release tag | https://github.com/onnx/onnx/issues/3084 | https://github.com/onnx/onnx/pull/3105 | 356108796c72f151fb4980e8c377b766265e76e4 | 50cea73255696c7d8b9013ffa38e46338716078f | 2020-11-02T09:24:50Z | python | 2022-04-06T21:08:11Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,017 | [".azure-pipelines/Linux-CI.yml", ".azure-pipelines/Windows-CI.yml", ".github/workflows/manylinux/entrypoint.sh", ".github/workflows/release_win.yml", ".github/workflows/weekly_mac_ci.yml", ".github/workflows/win_no_exception_ci.yml", "CMakeLists.txt", "README.md", "cmake/summary.cmake"] | Wrong Installation Guide in README.md | # Bug Report
### Describe the bug
The installation guide in https://github.com/onnx/onnx#linux-and-macos is wrong :|
It should tell users how to build onnx from source, but `pip install onnx` there will directly install the pre-compiled onnx package from PyPI.
In my opinion, a `--no-binary` command argument is needed https://pip.pypa.io/en/stable/reference/pip_install/#cmdoption-no-binary
| https://github.com/onnx/onnx/issues/3017 | https://github.com/onnx/onnx/pull/3550 | 6c77a411215305c1a3f355f8c2e4633c19514690 | 31f0fd6103e231664d1cfd3b25264f672f5de960 | 2020-09-13T13:09:40Z | python | 2021-07-02T21:24:44Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 3,013 | ["onnx/defs/controlflow/defs.cc", "onnx/defs/controlflow/old.cc"] | Shape inference issue with Loop output | # Bug Report
### Is the issue related to model conversion? No
### Describe the bug
Reported by @BowenBao . Adding the issue here to be able to track it and reference in a PR.
[loop_repro.zip](https://github.com/onnx/onnx/files/5211869/loop_repro.zip)
Attached are 2 models loop_pass and loop_fail
Download and run the script
import onnxruntime
import torch
x = torch.randn(3,4,5)
y = torch.randn(4,5)
sess = onnxruntime.InferenceSession('model.onnx')
ort_out = sess.run(None, {sess.get_inputs()[0].name: x.numpy(), sess.get_inputs()[1].name: y.numpy()})
Observe the following error message
onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Node (Concat_33) Op (Concat) [ShapeInferenceError] axis must be in [-rank, rank-1].
However, inspecting the model it is obvious the input has rank 3 which makes the range valid for axis=2.
| https://github.com/onnx/onnx/issues/3013 | https://github.com/onnx/onnx/pull/3014 | 04357970e33cd3730c8b2b3e726fb470f323979b | b8f6254bf133cf894611eb0f0293720473cf3518 | 2020-09-11T23:42:18Z | python | 2020-09-14T22:03:27Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 2,991 | ["docs/Changelog.md", "docs/Operators.md", "onnx/defs/tensor/defs.cc"] | ScatterElements and ScatterND input 'updates' should be differentiable | as title. | https://github.com/onnx/onnx/issues/2991 | https://github.com/onnx/onnx/pull/3028 | 63369697867ada4c88881175edc081728f445570 | b2ed660d0a065b8346816f2c3a95d79ca79b88c9 | 2020-09-02T06:17:02Z | python | 2020-09-23T20:24:27Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 2,965 | ["onnx/defs/tensor/utils.cc", "onnx/shape_inference/implementation.cc", "onnx/test/shape_inference_test.py"] | Shape inference on the Torchvision’s Mask R-CNN causes a segmentation fault | # Bug Report
### Is the issue related to model conversion?
Probably not.
### Describe the bug
When I try to run shape inference on the Torchvision's Mask R-CNN it causes a segfault. However, `check_model` doesn't return any warning/exception.
### System information
- OS Platform and Distribution (*e.g. Linux Ubuntu 16.04*): Linux Ubuntu 18.04.4
- ONNX version (*e.g. 1.7*): 1.6.0
- Python version: 3.6.10
- PyTorch version: 1.6.0a0+9907a3e
- Torchvision version: 0.7.0a0
- GCC/Compiler version (if compiling from source): N/A
- CMake version: 3.14.0
- Protobuf version: 3.12.2
- Visual Studio version (if applicable): N/A
### Reproduction instructions
```
import torch
import torchvision
import onnx
model = torchvision.models.detection.maskrcnn_resnet50_fpn(pretrained=True)
model.eval()
x = [torch.rand(3, 800, 800)]
torch.onnx.export(model, x, "mask_rcnn.onnx", opset_version = 11)
onnx_model = onnx.load("mask_rcnn.onnx")
onnx.checker.check_model(onnx_model)
onnx.shape_inference.infer_shapes(onnx_model)
```
### Expected behavior
Shape inference doesn't cause any errors. | https://github.com/onnx/onnx/issues/2965 | https://github.com/onnx/onnx/pull/2983 | 1be9a7de8f720cebbd794d596dc5cfc4d5b1859c | 53057782d19a59548705802fca756d87dee032e2 | 2020-08-19T11:40:19Z | python | 2020-08-28T23:19:50Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 2,481 | ["CMakeLists.txt"] | Build broken: Unknown generator option: dllexport_decl | @orionr @linkerzhang @zrphercule
Run on macOS Catalina 10.15.1 with protobuf 3.11.0:
```
git clone https://github.com/onnx/onnx.git && cd onnx
git submodule update --init --recursive
python3 setup.py install
```
Log:
```
~/onnx: python3 setup.py install
running install
running bdist_egg
running egg_info
creating onnx.egg-info
writing onnx.egg-info/PKG-INFO
writing dependency_links to onnx.egg-info/dependency_links.txt
writing entry points to onnx.egg-info/entry_points.txt
writing requirements to onnx.egg-info/requires.txt
writing top-level names to onnx.egg-info/top_level.txt
writing manifest file 'onnx.egg-info/SOURCES.txt'
reading manifest file 'onnx.egg-info/SOURCES.txt'
reading manifest template 'MANIFEST.in'
writing manifest file 'onnx.egg-info/SOURCES.txt'
installing library code to build/bdist.macosx-10.15-x86_64/egg
running install_lib
running build_py
running create_version
running cmake_build
-- The C compiler identification is AppleClang 11.0.0.11000033
-- The CXX compiler identification is AppleClang 11.0.0.11000033
-- Check for working C compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/cc
-- Check for working C compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++
-- Check for working CXX compiler: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found Protobuf: /usr/local/lib/libprotobuf.dylib (found version "3.11.0")
Generated: /Users/x/onnx/.setuptools-cmake-build/onnx/onnx-ml.proto
Generated: /Users/x/onnx/.setuptools-cmake-build/onnx/onnx-operators-ml.proto
CMake Warning at CMakeLists.txt:406 (find_package):
By not providing "Findpybind11.cmake" in CMAKE_MODULE_PATH this project has
asked CMake to find a package configuration file provided by "pybind11",
but CMake did not find one.
Could not find a package configuration file provided by "pybind11"
(requested version 2.2) with any of the following names:
pybind11Config.cmake
pybind11-config.cmake
Add the installation prefix of "pybind11" to CMAKE_PREFIX_PATH or set
"pybind11_DIR" to a directory containing one of the above files. If
"pybind11" provides a separate development package or SDK, be sure it has
been installed.
--
-- ******** Summary ********
-- CMake version : 3.15.5
-- CMake command : /usr/local/Cellar/cmake/3.15.5/bin/cmake
-- System : Darwin
-- C++ compiler : /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/c++
-- C++ compiler version : 11.0.0.11000033
-- CXX flags : -Wnon-virtual-dtor
-- Build type : Release
-- Compile definitions :
-- CMAKE_PREFIX_PATH :
-- CMAKE_INSTALL_PREFIX : /usr/local
-- CMAKE_MODULE_PATH :
--
-- ONNX version : 1.6.0
-- ONNX NAMESPACE : onnx
-- ONNX_BUILD_TESTS : OFF
-- ONNX_BUILD_BENCHMARKS : OFF
-- ONNX_USE_LITE_PROTO : OFF
-- ONNXIFI_DUMMY_BACKEND : OFF
-- ONNXIFI_ENABLE_EXT : OFF
--
-- Protobuf compiler : /usr/local/bin/protoc
-- Protobuf includes : /usr/local/include
-- Protobuf libraries : /usr/local/lib/libprotobuf.dylib
-- BUILD_ONNX_PYTHON : ON
-- Python version :
-- Python executable : /usr/local/opt/python/bin/python3.7
-- Python includes : /usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/include/python3.7m
-- Configuring done
-- Generating done
-- Build files have been written to: /Users/x/onnx/.setuptools-cmake-build
Scanning dependencies of target gen_onnx_proto
Scanning dependencies of target onnxifi_loader
Scanning dependencies of target onnxifi_dummy
[ 1%] Running gen_proto.py on onnx/onnx.in.proto
[ 4%] Building C object CMakeFiles/onnxifi_loader.dir/onnx/onnxifi_loader.c.o
[ 4%] Building C object CMakeFiles/onnxifi_dummy.dir/onnx/onnxifi_dummy.c.o
/Users/x/onnx/onnx/onnxifi_dummy.c:173:21: warning: incompatible function pointer types assigning to 'onnxExtensionFunctionPointer'
(aka 'int (*)(void)') from 'onnxStatus (*)(onnxBackendID, const char *, onnxExtensionFunctionPointer *)' (aka
'int (*)(void *, const char *, int (**)(void))') [-Wincompatible-function-pointer-types]
*function = &onnxGetExtensionFunctionAddress;
^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/Users/x/onnx/onnx/onnxifi_dummy.c:176:21: warning: incompatible function pointer types assigning to 'onnxExtensionFunctionPointer'
(aka 'int (*)(void)') from 'onnxStatus (*)(onnxGraph, uint32_t, const onnxTensorDescriptorV1 *, uint32_t, const onnxTensorDescriptorV1 *,
onnxMemoryFenceV1 *)' (aka 'int (*)(void *, unsigned int, const struct onnxTensorDescriptorV1 *, unsigned int, const struct onnxTensorDescriptorV1 *,
struct onnxMemoryFenceV1 *)') [-Wincompatible-function-pointer-types]
*function = &onnxSetIOAndRunGraph;
^ ~~~~~~~~~~~~~~~~~~~~~
Processing /Users/x/onnx/onnx/onnx.in.proto
Writing /Users/x/onnx/.setuptools-cmake-build/onnx/onnx-ml.proto
Writing /Users/x/onnx/.setuptools-cmake-build/onnx/onnx-ml.proto3
generating /Users/x/onnx/.setuptools-cmake-build/onnx/onnx_pb.py
2 warnings generated.
[ 6%] Linking C static library libonnxifi_loader.a
[ 9%] Linking C shared library libonnxifi_dummy.dylib
[ 9%] Running C++ protocol buffer compiler on /Users/x/onnx/.setuptools-cmake-build/onnx/onnx-ml.proto
[ 9%] Built target onnxifi_loader
--python_out: onnx/onnx-ml.proto: Unknown generator option: dllexport_decl
make[2]: *** [onnx/onnx-ml.pb.cc] Error 1
make[1]: *** [CMakeFiles/gen_onnx_proto.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
Scanning dependencies of target onnxifi_wrapper
[ 9%] Built target onnxifi_dummy
[ 11%] Building C object CMakeFiles/onnxifi_wrapper.dir/onnx/onnxifi_wrapper.c.o
[ 13%] Linking C shared module libonnxifi.dylib
[ 13%] Built target onnxifi_wrapper
make: *** [all] Error 2
Traceback (most recent call last):
File "setup.py", line 336, in <module>
'backend-test-tools = onnx.backend.test.cmd_tools:main',
File "/usr/local/lib/python3.7/site-packages/setuptools/__init__.py", line 145, in setup
return distutils.core.setup(**attrs)
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/core.py", line 148, in setup
dist.run_commands()
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/dist.py", line 966, in run_commands
self.run_command(cmd)
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/dist.py", line 985, in run_command
cmd_obj.run()
File "/usr/local/lib/python3.7/site-packages/setuptools/command/install.py", line 67, in run
self.do_egg_install()
File "/usr/local/lib/python3.7/site-packages/setuptools/command/install.py", line 109, in do_egg_install
self.run_command('bdist_egg')
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/dist.py", line 985, in run_command
cmd_obj.run()
File "/usr/local/lib/python3.7/site-packages/setuptools/command/bdist_egg.py", line 172, in run
cmd = self.call_command('install_lib', warn_dir=0)
File "/usr/local/lib/python3.7/site-packages/setuptools/command/bdist_egg.py", line 158, in call_command
self.run_command(cmdname)
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/dist.py", line 985, in run_command
cmd_obj.run()
File "/usr/local/lib/python3.7/site-packages/setuptools/command/install_lib.py", line 11, in run
self.build()
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/command/install_lib.py", line 105, in build
self.run_command('build_py')
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/dist.py", line 985, in run_command
cmd_obj.run()
File "setup.py", line 213, in run
self.run_command('cmake_build')
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/distutils/dist.py", line 985, in run_command
cmd_obj.run()
File "setup.py", line 207, in run
subprocess.check_call(build_args)
File "/usr/local/Cellar/python/3.7.5/Frameworks/Python.framework/Versions/3.7/lib/python3.7/subprocess.py", line 363, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['/usr/local/bin/cmake', '--build', '.', '--', '-j', '16']' returned non-zero exit status 2.
``` | https://github.com/onnx/onnx/issues/2481 | https://github.com/onnx/onnx/pull/2482 | 477a9b87715d614f8b7540a69c144b177275baa2 | 57ebc587fcf3913b4be93653b0dd58c686447298 | 2019-11-29T19:59:12Z | python | 2019-12-17T01:01:14Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 2,449 | ["docs/Changelog.md", "docs/Operators.md", "onnx/defs/tensor/defs.cc", "onnx/defs/tensor/old.cc", "onnx/version_converter/adapters/softmax_12_13.h"] | Spec unclear: Can a scalar be Reshaped to a 1D tensor and vice versa? | A scalar has shape [ ] and rank 0. Can the Reshape operator be applied to it with a shape of [ 1 ]? And Can a 1D tensor of shape [ 1 ] have Reshape applied with a shape of [ ]? | https://github.com/onnx/onnx/issues/2449 | https://github.com/onnx/onnx/pull/3861 | ae3e6b80e63fbef0abda96ce698d31c70cd5e43d | dba3bd832f4d740d991d367535bab5c3634455a7 | 2019-11-12T17:59:26Z | python | 2022-01-25T01:20:53Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 2,446 | ["onnx/optimizer/passes/fuse_consecutive_transposes.h", "onnx/test/optimizer_test.py"] | onnx optimizer cannot fuse consecutive transposes correctly. | The following is a simple repro code:
```
import onnx
from onnx import helper, TensorProto
from onnx import optimizer, shape_inference
# Preprocessing: create a model with two nodes, Y's shape is unknown
node1 = helper.make_node('Transpose', ['X'], ['Y'], perm=[1, 0, 2])
node2 = helper.make_node('Transpose', ['Y'], ['Z'], perm=[0, 2, 1])
graph = helper.make_graph(
[node1, node2],
'two-transposes',
[helper.make_tensor_value_info('X', TensorProto.FLOAT, (2, 3, 4))],
[helper.make_tensor_value_info('Z', TensorProto.FLOAT, (3, 4, 2))],
)
original_model = helper.make_model(graph, producer_name='onnx-examples')
# Apply shape inference on the model and check the it
inferred_model = shape_inference.infer_shapes(original_model)
onnx.checker.check_model(inferred_model)
# Pick one pass as example
passes = ['fuse_consecutive_transposes']
# Apply the optimization on the original mode and check it
optimized_model = optimizer.optimize(original_model, passes)
onnx.checker.check_model(shape_inference.infer_shapes(optimized_model)) #error happens here in infer_shapes
print('The model after optimization:\n{}'.format(optimized_model))
```
The optimized model fails to pass infer_shapes.
I looked into it. The 2 consecutive transpose ops along axis [1,0,2] and [0, 2,1] should be fused to [1,2,0]. But onnx gives a fused one [2,0,1].
| https://github.com/onnx/onnx/issues/2446 | https://github.com/onnx/onnx/pull/2471 | ad1f5567b6c00cfde71a2c0e466b07101d8c84c2 | cdc8b86196ad4d20e9e7e530b3323f23324babea | 2019-11-12T07:15:22Z | python | 2019-12-05T06:04:00Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 2,339 | ["third_party/pybind11"] | import paddle.fluid & import onnx will be core_dump | https://github.com/microsoft/onnxruntime/issues/1541
https://github.com/pybind/pybind11/issues/1262
Update pybind to the latest version can solve this. | https://github.com/onnx/onnx/issues/2339 | https://github.com/onnx/onnx/pull/2340 | 7840504d4d6f3125f6ae4dff256381632cf84d4f | 1aa176e05b722b408b83aaf2b2a893c9fbaa37d7 | 2019-09-19T17:27:48Z | python | 2019-09-19T20:35:31Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 2,333 | ["docs/Changelog.md", "docs/Operators.md", "onnx/defs/nn/defs.cc"] | ConvTranspose's output_padding attribute requires clarification | ConvTranspose's output_padding attribute is defined as follows:
> The **zero-padding** added to one side of the output. This is also called adjs/adjustment in some frameworks.
(Emphasis mine.) The most likely interpretation of this is that a positive output_padding would add rows/columns of zeros to one side of the output tensor. This is misleading, as the apparent intention of this attribute is to match PyTorch's definition, which is different. PyTorch instead defines it as follows:
https://pytorch.org/docs/stable/nn.html#convtranspose2d
> The padding argument effectively adds dilation * (kernel_size - 1) - padding amount of zero padding to both sizes of the input. This is set so that when a Conv2d and a ConvTranspose2d are initialized with same parameters, they are inverses of each other in regard to the input and output shapes. However, when stride > 1, Conv2d maps multiple input shapes to the same output shape. output_padding is provided to resolve this ambiguity by effectively increasing the calculated output shape on one side. **Note that output_padding is only used to find output shape, but does not actually add zero-padding to output.**
(Emphasis mine.) In other words in PyTorch, output_padding affects the output shape (just like it does in ONNX) but the added rows/columns *are not zero'd* - they participate in the ConvTranspose just like the other elements of the output tensor.
The definition of output_padding in ConvTranspose should therefore probably be clarified to say something closer to how PyTorch defines it - that is, that it participates in shape inference of the output tensor, but doesn't actually zero out those added rows/columns.
(This was discussed earlier in an offline thread with @spandantiwari and @pranavsharma) | https://github.com/onnx/onnx/issues/2333 | https://github.com/onnx/onnx/pull/2343 | a20ba2f1dc3fbd1109fcc67bf42247531832bb60 | c5af774aa950f6a5cf1a7b5f37d014fb2fbd82bc | 2019-09-18T22:22:40Z | python | 2019-09-21T00:52:48Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,949 | ["onnx/defs/tensor/defs.cc", "onnx/defs/tensor_proto_util.cc", "onnx/defs/tensor_proto_util.h", "onnx/test/shape_inference_test.py"] | shape inference bug of op "Slice 10" | this shape inference function fails even in simple cases, for example a ONNX graph that only contains 1 slice whose inputs such as "starts" are all const nodes except the "data".
| https://github.com/onnx/onnx/issues/1949 | https://github.com/onnx/onnx/pull/1950 | 6fb0775d85226668048360cacba4763deb80d948 | 414dbc731506e21c70b4600352a161c528233a57 | 2019-04-18T03:10:59Z | python | 2019-04-19T07:35:37Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,934 | ["docs/IR.md"] | Case sensitivity of operator names and attributes | The spec should be clear about case sensitivity when matching names of operators, attributes, and attribute string enums. That includes NodeProto::op_type, AttributeProto::name, and AttributeProto::s/strings.
## Applicable documentation:
https://github.com/onnx/onnx/blob/master/docs/IR.md
https://github.com/onnx/onnx/blob/master/docs/Operators.md#RNN
https://github.com/onnx/onnx/blob/master/onnx/onnx.proto
## Context:
We’ve been matching all of these in WinML kernels atop DirectML with exact case sensitivity (simpler/faster/less ambiguous), but we encountered a failing model exported from PyTorch where the RNN activation attribute “Tanh” was “tanh” instead :/. Granted, I see this as a bug in the exporter because the exporter should use the same case as the spec lists them, but the spec should make it clear in any case whether "tanh" should match too. | https://github.com/onnx/onnx/issues/1934 | https://github.com/onnx/onnx/pull/4096 | 61c36aa4250ca607c194fc55070e95db2b95761b | ee4888c24510787bb8c61ebcba43ef252744e648 | 2019-04-12T21:55:54Z | python | 2022-03-29T15:30:28Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,823 | ["onnx/backend/test/case/model/shrink.py", "onnx/backend/test/data/node/test_shrink_hard/model.onnx", "onnx/backend/test/data/node/test_shrink_soft/model.onnx", "onnx/backend/test/data/simple/test_shrink/model.onnx"] | test data for shrink_test is of incorrect shape | Both onnx/backend/test/data/simple/test_shrink/test_data_set_0/input_0.pb and onnx/backend/test/data/simple/test_shrink/test_data_set_0/output_0.pb have shape [5] whereas the graph input and output in onnx/backend/test/data/simple/test_shrink/model.onnx has shape [1,5].
In test_shrink_soft and test_shrink_hard, the test input data, test output data, the graph input defined in the model, the graph output defined in the model all have consistent shape of [5].
Tagging @zrphercule as fyi | https://github.com/onnx/onnx/issues/1823 | https://github.com/onnx/onnx/pull/1825 | b37fc6df4e426650cb0d48fc69519d0cae1584ae | 80346bd748fc8e586c70e89c50befe7d7f8c6383 | 2019-02-19T22:33:37Z | python | 2019-02-20T22:54:18Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,775 | ["docs/Changelog.md", "docs/Operators.md", "onnx/defs/logical/defs.cc"] | doc issue for Greater op | in Opset9, it is not limited to ```float tensors``` anymore.
https://github.com/onnx/onnx/blob/15c33c945851907411619f599900c3852108e7e3/onnx/defs/logical/defs.cc#L82-L94 | https://github.com/onnx/onnx/issues/1775 | https://github.com/onnx/onnx/pull/1811 | 0628582506ede89f94faa80977d86e3f7e19739c | 4f064a16359f256d8f7ba13f957a9013451c319a | 2019-01-29T06:58:25Z | python | 2019-02-14T16:02:58Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,755 | ["docs/Changelog.md", "docs/Operators.md", "onnx/defs/tensor/defs.cc"] | Inaccurate doc for Squeeze | For attribute `axes` in `Squeeze`, doc describe it as:
```
List of positive integers, indicate the dimensions to squeeze.
```
But based on the implementation in Caffe2 and Microsoft's onnxruntime, this is 0-based instead of positive ints. | https://github.com/onnx/onnx/issues/1755 | https://github.com/onnx/onnx/pull/1905 | a8b45d62de57ab9b25d7764e38e44f5e9e9c027a | 079c2639f9bb79b1774d1e3bfa05b0c093816ca7 | 2019-01-21T20:28:38Z | python | 2019-04-03T06:21:37Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,735 | ["onnx/defs/tensor/defs.cc", "onnx/test/shape_inference_test.py"] | Shape inference fails with a Split node with an split attribute | When a `split` attribute is set to a `Split` node, `onnx.shape_inference.infer_shapes` fails to infer its output shapes.
```
import onnx
import onnx.shape_inference
op = onnx.helper.make_node("Split", inputs=["a"], outputs=["b", "c", "d"], axis=0)
i = onnx.helper.make_tensor_value_info(name="a", elem_type=onnx.TensorProto.FLOAT, shape=[6])
model = onnx.helper.make_model(onnx.helper.make_graph([op], "graph", [i], []))
print(onnx.shape_inference.infer_shapes(model)) # contains proper value_info
op = onnx.helper.make_node("Split", inputs=["a"], outputs=["b", "c", "d"], axis=0, split=[1, 2, 3])
i = onnx.helper.make_tensor_value_info(name="a", elem_type=onnx.TensorProto.FLOAT, shape=[6])
model = onnx.helper.make_model(onnx.helper.make_graph([op], "graph", [i], []))
print(onnx.shape_inference.infer_shapes(model)) # the shapes of b, c, and d are not inferred
```
While the former example outputs the correct shapes (three `[2]`), the resulting `value_info` of the latter example do not contain dimension information.
This will happen even if the tensor `a` is divided evenly (by setting `split=[2, 2, 2]`).
I'm using ONNX 1.3.0. | https://github.com/onnx/onnx/issues/1735 | https://github.com/onnx/onnx/pull/2328 | 684fc1bcf29da1986ffa5a9c33bef0565bb1a894 | dfa4384c3f8fe9d804032f992e6f818fc54152b0 | 2019-01-14T22:01:07Z | python | 2020-01-14T23:57:28Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,689 | ["docs/Changelog.md", "docs/Operators.md", "onnx/defs/experiments/defs.cc", "onnx/defs/gen_doc.py", "onnx/defs/operator_sets.h"] | Either DynamicSlice is experimental or part of opset 9. Which one is it? | The op is classified under [experimental][1] while it's [mentioned][2] that it's part of opset 9.
[1]: https://github.com/onnx/onnx/blob/master/docs/Operators.md#experimental-dynamicslice
[2]: https://github.com/onnx/onnx/blob/master/docs/Operators.md#version-122 | https://github.com/onnx/onnx/issues/1689 | https://github.com/onnx/onnx/pull/1696 | 0c8d857bb162431912b255d5c0e773fb7c131a65 | da7c50cb0e5948e43b9d7aa72489b1e040090461 | 2018-12-12T03:06:08Z | python | 2018-12-19T15:53:31Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,661 | [".github/workflows/codeql.yml", ".github/workflows/lint.yml", ".github/workflows/scorecard.yml"] | Specifications for SVMClassifier is not precise enough | There are N columns for N classes if the operators output probabilities but if N > 2. What happens when the operator outputs raw scores for N(N-1)/2 support vectors? libsvm returns N(N-1)/2 columns, scikit-learn normalizes the scores to get N columns. The spec should choose one behavior or add a parameter to specify which one is expected.
| https://github.com/onnx/onnx/issues/1661 | https://github.com/onnx/onnx/pull/5186 | b662b2c04dd1ef665132a98ed27c2c32c35d4643 | ee593e5af421e93770dd454f68a5154793c9c7ae | 2018-11-29T16:20:29Z | python | 2023-05-01T17:05:24Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,564 | ["onnx/common/interned_strings.h", "onnx/examples/optimize_onnx.ipynb", "onnx/optimizer/pass_registry.h", "onnx/optimizer/passes/eliminate_nop_dropout.h", "onnx/optimizer/passes/fuse_consecutive_reduce_unsqueeze.h", "onnx/test/optimizer_test.py"] | Optimization: Fuse {Reduction[keepdim=0]->Unsqueeze} to {Reduction[keepdim=1]} | https://github.com/onnx/onnx/issues/1564 | https://github.com/onnx/onnx/pull/1565 | 6db386e4ccd1a63a367489e5ebaf4433fc3204e3 | ea694bf5cbccf34e08f36a8d9365d9011b72aeb4 | 2018-10-31T01:28:41Z | python | 2018-11-20T08:22:25Z |
|
closed | onnx/onnx | https://github.com/onnx/onnx | 1,556 | ["onnx/common/interned_strings.h", "onnx/examples/optimize_onnx.ipynb", "onnx/optimizer/pass_registry.h", "onnx/optimizer/passes/fuse_consecutive_concats.h", "onnx/test/optimizer_test.py"] | Optimization Pass: Fusing consecutive concats | I was looking at an exported model at work and saw the following behavior:
In this case we can replace the 3 concats with a single concat along axis=0. I saw this type of pattern in multiple places of the DAG and it seems to be a biproduct of pytorch exportation of GRU (or maybe all RNN) ops. | https://github.com/onnx/onnx/issues/1556 | https://github.com/onnx/onnx/pull/1557 | 3265c1e1d953b2638059ddcd55cfc8dee4c24cd7 | 736dc863dbf002af2509dad2f030dd864ab2a196 | 2018-10-26T17:11:54Z | python | 2018-10-30T06:41:42Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,521 | ["onnx/common/interned_strings.h", "onnx/examples/optimize_onnx.ipynb", "onnx/optimizer/pass_registry.h", "onnx/optimizer/passes/eliminate_nop_dropout.h", "onnx/test/optimizer_test.py"] | Optimization Pass: Dropout with ratio 0.0 should be a Nop | https://github.com/onnx/onnx/issues/1521 | https://github.com/onnx/onnx/pull/1555 | f958d5a8f8581f5a1cf362724ef9f33054ca5ff4 | 3265c1e1d953b2638059ddcd55cfc8dee4c24cd7 | 2018-10-18T00:43:08Z | python | 2018-10-29T05:20:41Z |
|
closed | onnx/onnx | https://github.com/onnx/onnx | 1,429 | ["onnx/common/interned_strings.h", "onnx/examples/optimize_onnx.ipynb", "onnx/optimizer/pass_registry.h", "onnx/optimizer/passes/eliminate_nop_monotone_argmax.h", "onnx/test/optimizer_test.py"] | Optimization Pass: Monotonic growing function is useless before argmax calculation. | It's standard to use softmax to generate a probability distribution at the end of a model. If we want to get the predicted class we end up with `net(x).softmax(dim=1).argmax(dim=1)`.
In reality though because softmax is a monotonic growing operation we can replace the expression with `net(x).argmax(dim=1)` | https://github.com/onnx/onnx/issues/1429 | https://github.com/onnx/onnx/pull/1519 | 7b7b6ec4fdf21039c1e77814d1a1cb7c2d658e9a | f958d5a8f8581f5a1cf362724ef9f33054ca5ff4 | 2018-09-19T21:46:42Z | python | 2018-10-26T04:27:30Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,313 | ["onnx/backend/test/case/node/mvn.py", "onnx/backend/test/data/node/test_mvn/model.onnx", "onnx/backend/test/data/node/test_mvn/test_data_set_0/input_0.pb", "onnx/backend/test/data/node/test_mvn/test_data_set_0/output_0.pb"] | Test Cases for FuncMeanVarianceNormalization Missing | Test cases for FuncMeanVarianceNormalization are missing. This means that, on a separate branch of ONNX, when test data is generated, relevant tests start failing (schema for FuncMeanVarianceNormalization can't be found). | https://github.com/onnx/onnx/issues/1313 | https://github.com/onnx/onnx/pull/1335 | e509836f45ebbb7b9b977405d411eb042e24163d | ddcaf8adc3ed0d6c40d5c2691ad231593c40df61 | 2018-08-22T17:37:06Z | python | 2018-08-28T17:24:12Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,194 | ["CMakeLists.txt", "onnx/__init__.py", "onnx/onnx-operators-ml.proto", "onnx/onnx-operators-ml.proto3", "onnx/onnx-operators.in.proto", "onnx/onnx-operators.proto", "onnx/onnx-operators.proto3", "tools/protoc-gen-mypy.py"] | ModuleNotFoundError in generated onnx_operators_pb2.py | Need to import a class defined in operators.proto. Try adding `from .onnx_operators_pb import *` in __init__.py and had this error reported as follows:
```
>>> import onnx
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "C:\\onnx\onnx\__init__.py", line 8, in <module>
from .onnx_operators_pb import *
File "C:\\onnx\onnx\onnx_operators_pb.py", line 8, in <module>
from .onnx_operators_pb2 import * # noqa
File "C:\\onnx\onnx\onnx_operators_pb2.py", line 17, in <module>
import onnx_pb2 as onnx__pb2
ModuleNotFoundError: No module named 'onnx_pb2'
```
if I manually change ` import onnx_pb2 as onnx__pb2` to ` import onnx.onnx_pb2 as onnx__pb2` it wont crash. Any thoughts on how to bypass it? | https://github.com/onnx/onnx/issues/1194 | https://github.com/onnx/onnx/pull/1198 | 17818ea03ae575e5c0777d69402918ce2f623902 | 2e7099ee7c37b196c197c9a084a97698a41da232 | 2018-07-13T00:49:39Z | python | 2018-07-16T22:30:13Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 1,100 | ["docs/IR.md"] | Missing/wrong documentation for opset_import/OperatorSetId | The OperatorSetId is not defined in https://github.com/onnx/onnx/blob/master/docs/IR.md
> opset_import | OperatorSetId | A collection of operator set identifiers made available to the model. An implementation must support all operators in the set or reject the model.
`operator set identifiers` may be assumed to be a simple string. In fact it is a `{domain, version}` struct. | https://github.com/onnx/onnx/issues/1100 | https://github.com/onnx/onnx/pull/4039 | c940fa3fea84948e46603cab2f86467291443beb | 5a6628799ee4fa31edc4451886363f6d47095caa | 2018-06-08T12:07:17Z | python | 2022-02-28T20:43:55Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 859 | [".github/workflows/stale.yml"] | Request for adding Pack/Unpack to onnx (that are the ops tf.stack/tf.unstack) | After the discussion with @guschmue, we think adding Pack/Unpack to onnx is necessary for implementing Strideslices op of Tensorflow. I wish tf.stack/tf.unstack could be added to onnx | https://github.com/onnx/onnx/issues/859 | https://github.com/onnx/onnx/pull/4917 | 5849e3ed84fad5fbfe18215a0a053f1fcbea47b4 | dad6338fc2bf28c353d7c797122d9b63190e8495 | 2018-04-30T21:59:43Z | python | 2023-02-21T14:53:32Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 708 | ["cmake/Utils.cmake", "setup.py"] | CMake setup fail on local dev build | See error from cmake when build using `pip install -e .` . Looking into it now
```
ERROR Failed to remove indentation from:
"""
from distutils import sysconfig
print(sysconfig.get_python_lib(prefix=''))
"""
Python dedent failed with error code: The system cannot find the file specified
CMake Error at cmake/Utils.cmake:31 (message):
Python dedent failed with error code: The system cannot find the file specified
``` | https://github.com/onnx/onnx/issues/708 | https://github.com/onnx/onnx/pull/709 | 6734224600c78621e516e50a1e17fd1cbab9cb0b | 93f0d401612ee76afc2924fee0e14a01e27662f8 | 2018-04-04T17:27:45Z | python | 2018-04-04T20:13:31Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 518 | [".github/workflows/stale.yml"] | torch.onnx | When I execute ‘python -c" import torch.onnx "‘
`bainuo@bainuo-SYS-7048GR-TR:~$ python -c "import torch.onnx"
Traceback (most recent call last):
File "<string>", line 1, in <module>
ImportError: No module named onnx`
What should I do | https://github.com/onnx/onnx/issues/518 | https://github.com/onnx/onnx/pull/4917 | 5849e3ed84fad5fbfe18215a0a053f1fcbea47b4 | dad6338fc2bf28c353d7c797122d9b63190e8495 | 2018-02-07T03:12:30Z | python | 2023-02-21T14:53:32Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 455 | [".github/workflows/stale.yml"] | [PR 400 Follow Up] Tune the IR API | Further discussion is needed. | https://github.com/onnx/onnx/issues/455 | https://github.com/onnx/onnx/pull/4917 | 5849e3ed84fad5fbfe18215a0a053f1fcbea47b4 | dad6338fc2bf28c353d7c797122d9b63190e8495 | 2018-01-24T06:43:37Z | python | 2023-02-21T14:53:32Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 405 | [".github/workflows/stale.yml"] | Gemm requires (MxK), (KxN) inputs but no reshape are inserted in onnx models. | I facing a problem when working with "bvlc_alexnet/model.pb" from [here](https://github.com/onnx/models/tree/master/bvlc_alexnet)
Its include a *Gemm* node with 3 inputs:
- The output of the previous node : pool_5, its shape is : [1, 256, 6, 6]
- A weights tensor presents into the initializers. its shape is : [4096, 9216]
- A bias tensor presents into initializers. its shape is : [4096]
The attrs of the node are :
- Broadcasting = 1 : it's ok to broadcast the bias tensor.
- transB = 1 : it's ok to transpose the weights tensor.
In the [doc](https://github.com/onnx/onnx/blob/master/docs/Operators.md#gemm) we can read :
> where input tensor A has dimension (M X K) , input tensor B has dimension (K X N), input tensor C and output tensor Y have dimension (M X N)
So the tensor B will have shape : [9216, 4096] after being transposed.
Si the input A (the output of the previous node) MUST have the shape : [1, 256\*6\*6] = [1, 9216]
This will produce a tensor of shape [1, 4096] which we can add to the bias tensor (input C).
According to me, a reshape is missing between the pool_5 node and the Gemm node.
| https://github.com/onnx/onnx/issues/405 | https://github.com/onnx/onnx/pull/4917 | 5849e3ed84fad5fbfe18215a0a053f1fcbea47b4 | dad6338fc2bf28c353d7c797122d9b63190e8495 | 2018-01-09T14:08:26Z | python | 2023-02-21T14:53:32Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 374 | ["docs/Changelog.md", "docs/Operators.md", "onnx/defs/schema.h", "onnx/defs/tensor/defs.cc", "onnx/defs/tensor/old.cc"] | Concat Axis | Unless my protobuf implementation is up the spout (which is possible) the squeezenet model has no axis parameter in the Concat nodes. There is no default indicated. Squeezenet should be concatting on channels. What does no axis parameter indicate since the docs show no default? | https://github.com/onnx/onnx/issues/374 | https://github.com/onnx/onnx/pull/390 | 938ce339fd3e7dc5eb6a8f6332f77762ff5eaaaf | 535c8c1ded151c772051dd8f7a0887752c2d4fd5 | 2017-12-14T22:42:31Z | python | 2018-02-20T22:43:32Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 263 | [".github/workflows/auto_update_doc.yml"] | Remove pip install dependency on protoc | When the ONNX Python package becomes a dependency of a project, then currently the Protocol Buffers compiler becomes a dependency.
When installing the ONNX Python package many users get the following error:
```
Could not find "protoc" executable!
```
This is caused by the missing Protocol Buffers compiler. On some systems it's easy to install Protocol Buffers, but on Centos or Ubuntu 14.04 this is not so simple. Packaged versions of Protocol Buffers are older than 2.6.1, so users get these errors:
```
onnx.proto:308:7: Expected "required", "optional", or "repeated".
onnx.proto:308:19: Missing field number.
onnx.proto:330:3: Expected "required", "optional", or "repeated".
onnx.proto:330:15: Missing field number.
```
Would it be possible to remove the dependency on `protoc` from the `pip install` process?
Could the PyPI package be distributed with a precompiled form of the proto?
| https://github.com/onnx/onnx/issues/263 | https://github.com/onnx/onnx/pull/5730 | 4b7049dae24a8d233a66d2a378aafd7442896bd3 | f97f16fbb41bbf3b1a1ce5e063f9d8773ae342b8 | 2017-11-15T16:43:10Z | python | 2023-11-01T19:04:54Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 91 | [".github/workflows/clang_tidy_review.yml", ".github/workflows/clang_tidy_review_post.yml"] | Reference exports (Model zoo) | It may be a good idea to add a set of models serialized in the `onnx.proto` format to the `onnx/examples` directory. This could serve as a reference for people testing ONNX or writing their own import/export functions. What do you think?
I would propose starting with a set of very simple models, getting progressively more and more complex: a simple addition (a+b+c), linear regression, logistic regression, MLP and then more advanced models such as AlexNet, VGG, ResNet, etc.
| https://github.com/onnx/onnx/issues/91 | https://github.com/onnx/onnx/pull/5503 | 329970e15867fe830377ec43bb562c90f1892481 | e90f9f2d051fac6f299ff32fcb218a635dd76615 | 2017-10-09T10:43:56Z | python | 2023-08-14T16:11:18Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 90 | [".github/workflows/clang_tidy_review.yml", ".github/workflows/clang_tidy_review_post.yml"] | Table of contents for operators | It can be sometimes tricky to jump to the definition of an operator like "Pad", where the string "pad" occurs multiple times in Operators.md. A modest usability improvement would be to add a table of contents, so that string search first discovers operator names, before a full text search. | https://github.com/onnx/onnx/issues/90 | https://github.com/onnx/onnx/pull/5478 | 584624997b7e07b3bc80e7b437e04c124399f379 | 48f28c3eb6ebdb70e010d92ce1f33d987cf56dce | 2017-10-08T15:44:56Z | python | 2023-08-07T16:27:34Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 78 | ["onnx/defs/data_type_utils.cc", "onnx/onnx-ml.proto", "onnx/onnx-ml.proto3", "onnx/onnx.in.proto", "onnx/onnx.proto", "onnx/onnx.proto3"] | Protobuf naming consistency / nested protos | We've been keeping an eye on annoyances when we adjust PyTorch for latest changes in the ONNX proto, and we noticed some for https://github.com/onnx/onnx/pull/51
* Not all protobuf messages have the `Proto` suffix anymore; in particular, the nested protobuf `Dimension` doesn't have the suffix. This makes it look strange when compared with all the other protobuf messages. Can we make this consistent?
* Additionally, `TypeProto` defines nested protobufs. I wasn't sure how this was going to look in tooling, but we have discovered what happens with nanopb: you now get extremely ugly protobuf names like `TypeProto_TensorTypeProto`. This is especially ugly because `Type` is redundantly specified twice. Can we just put it at the top level of the protobuf? | https://github.com/onnx/onnx/issues/78 | https://github.com/onnx/onnx/pull/215 | 3a8c7988fa551ad8d21bdf14df8afe3884a7e9fd | 211336180e1da62752027a7926ba24b4b7ae5a11 | 2017-10-05T17:43:24Z | python | 2017-11-14T00:49:47Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 53 | [".github/workflows/pages.yml"] | ir_version, publisher_tag, and publisher_version should appear in as the initial fields in GraphProto | It's typical to put versioning fields at the beginning of a file/message format to allow implementations to efficiently sniff the version prior to parsing the entire buffer.
Doing so would a breaking change, but it's still quite early and we may have time.
| https://github.com/onnx/onnx/issues/53 | https://github.com/onnx/onnx/pull/5049 | 0cb435406195e1921343c675bee47bee60359c5d | 06536487cb5f7e863dc11598d4693f6d3a8848a7 | 2017-09-24T22:11:16Z | python | 2023-03-27T21:48:35Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 28 | ["requirements-lintrunner.txt"] | ImportError: No module named version | With the latest `__version__` support commits, if you `pip install -e onnx` from source:
```
>>> import onnx
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "onnx/__init__.py", line 9, in <module>
from .version import version as __version__
ImportError: No module named version
```
I think this is because we are not generating version.py for develop builds.
Fix here probably must be applied to onnx-caffe2 as well. | https://github.com/onnx/onnx/issues/28 | https://github.com/onnx/onnx/pull/5726 | fd699fbb4f69a0d2d12129013145c9f89a8c7eb0 | 4b7049dae24a8d233a66d2a378aafd7442896bd3 | 2017-09-15T20:43:34Z | python | 2023-11-01T18:40:46Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 27 | ["requirements-lintrunner.txt"] | Operator: Softmax | Requested in https://github.com/pytorch/pytorch/issues/2747 | https://github.com/onnx/onnx/issues/27 | https://github.com/onnx/onnx/pull/5726 | fd699fbb4f69a0d2d12129013145c9f89a8c7eb0 | 4b7049dae24a8d233a66d2a378aafd7442896bd3 | 2017-09-15T14:14:34Z | python | 2023-11-01T18:40:46Z |
closed | onnx/onnx | https://github.com/onnx/onnx | 26 | ["requirements-lintrunner.txt"] | Operator: SubConstant/AddConstant | Add/subtract a constant from all elements of a tensor
Note that MulConstant is handled as "Scale" atm. | https://github.com/onnx/onnx/issues/26 | https://github.com/onnx/onnx/pull/5332 | 2843c9e8ccd6757019952a0dad0efcc7b2dcd0c5 | 42e091c2e628d896043086d173c0dbf321b78ffc | 2017-09-15T00:55:04Z | python | 2023-06-20T21:32:30Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,492 | ["dubbo-rpc/dubbo-rpc-rest/src/main/java/org/apache/dubbo/rpc/protocol/rest/extension/resteasy/filter/ResteasyRequestContainerFilterAdapter.java"] | memory leak in rest | I got this error log:
```
2023-11-23 19:23:26.786 ERROR 27 --- [NettyServerWorker-thread-1] io.netty.util.ResourceLeakDetector : LEAK: ByteBuf.release() was not called before it's garbage-collected. See https://netty.io/wiki/reference-counted-objects.html for more information.
Recent access records:
Created at:
io.netty.buffer.PooledByteBufAllocator.newDirectBuffer(PooledByteBufAllocator.java:385)
io.netty.buffer.AbstractByteBufAllocator.directBuffer(AbstractByteBufAllocator.java:187)
io.netty.buffer.AbstractByteBufAllocator.directBuffer(AbstractByteBufAllocator.java:173)
io.netty.buffer.AbstractByteBufAllocator.buffer(AbstractByteBufAllocator.java:107)
org.apache.dubbo.rpc.protocol.rest.extension.resteasy.ResteasyContext.convertHttpRequestToContainerRequestContext(ResteasyContext.java:81)
org.apache.dubbo.rpc.protocol.rest.extension.resteasy.filter.ResteasyRequestContainerFilterAdapter.filter(ResteasyRequestContainerFilterAdapter.java:54)
org.apache.dubbo.rpc.protocol.rest.handler.NettyHttpHandler.executeFilters(NettyHttpHandler.java:131)
org.apache.dubbo.rpc.protocol.rest.handler.NettyHttpHandler.handle(NettyHttpHandler.java:82)
org.apache.dubbo.rpc.protocol.rest.netty.RestHttpRequestDecoder.lambda$decode$0(RestHttpRequestDecoder.java:70)
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
org.apache.dubbo.common.threadlocal.InternalRunnable.run(InternalRunnable.java:41)
java.lang.Thread.run(Thread.java:748)
```
the reason:
org.apache.dubbo.rpc.protocol.rest.extension.resteasy.filter.ResteasyRequestContainerFilterAdapter#filter creates a DubboPreMatchContainerRequestContext. DubboPreMatchContainerRequestContext's NettyHttpRequest field contains a ByteBuf.
ResteasyRequestContainerFilterAdapter#filter doesn't release the ByteBuf before exit.
solution:
call containerRequestContext.getHttpRequest().releaseContentBuffer() before exit. like
```
finally {
containerRequestContext.getHttpRequest().releaseContentBuffer();
}
```
| https://github.com/apache/dubbo/issues/13492 | https://github.com/apache/dubbo/pull/13514 | 75f740df0a80dd869d17b9ceb577c35de356c041 | 092a1bd619ece992521d469bfc761fb2d9da56d0 | 2023-12-13T09:34:04Z | java | 2023-12-20T13:06:23Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,443 | ["dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/TriRpcStatus.java", "dubbo-rpc/dubbo-rpc-triple/src/test/java/org/apache/dubbo/rpc/TriRpcStatusTest.java"] | Caused by: com.alibaba.fastjson2.JSONException: not support none serializable class org.apache.dubbo.rpc.TriRpcStatus | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.2.8
* fastjson version: 2.0.41
* Java version: 1.8
### Steps to reproduce this issue
TriRpcStatus 没有继承Serializable接口 当执行到cace result.setException(status.asException());的时候回报错：
Caused by: com.alibaba.fastjson2.JSONException: not support none serializable class org.apache.dubbo.rpc.TriRpcStatus
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
<!-- What do you expect from the above steps？-->
### Actual Behavior
<!-- What actually happens? -->
If there is an exception, please attach the exception trace:
```
Just put your stack trace here!
```
| https://github.com/apache/dubbo/issues/13443 | https://github.com/apache/dubbo/pull/13453 | 58fa12079578b7def49fe6c386c1bcce97f2123a | 3e1310bea880add83510855a4506595ca488245e | 2023-11-30T09:05:45Z | java | 2023-12-04T04:48:27Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,436 | ["dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/TripleInvoker.java"] | 3.2.7 invoking remote service in generic mode causes java.lang.ClassNotFoundException | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: `3.2.7`
* Operating System version: `Linux 4.19.91-009.ali4000.alios7.x86_64 #1 SMP Mon Jan 25 10:47:38 CST 2021 x86_64 x86_64 x86_64 GNU/Linux`
* Java version: `OpenJDK Runtime Environment (Alibaba 8.4.7) (build 1.8.0_152-b187) OpenJDK 64-Bit Server VM (Alibaba 8.4.7) (build 25.152-b187, mixed mode)`
### Steps to reproduce this issue
1. define generic reference config
```java
referenceConfig.setGeneric("true");
referenceConfig.setAsync(false);
genericService = referenceConfig.get();
```
2. do invocation
```java
// argTypes contain remote-defined-only class
genericService.$invoke(method, argTypes, args);
```
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
Succeed in invoking remote service in generic mode.
### Actual Behavior
Failed to invoke remote in generic mode, exception occurred in client side.
```text
Caused by: org.apache.dubbo.rpc.StatusRpcException: INTERNAL : Call aborted cause client exception
at org.apache.dubbo.rpc.TriRpcStatus.asException(TriRpcStatus.java:214) ~[dubbo-3.2.7.jar:3.2.7]
at org.apache.dubbo.rpc.protocol.tri.TripleInvoker.doInvoke(TripleInvoker.java:166) ~[dubbo-3.2.7.jar:3.2.7]
at org.apache.dubbo.rpc.protocol.AbstractInvoker.doInvokeAndReturn(AbstractInvoker.java:243) ~[dubbo-3.2.7.jar:3.2.7]
at org.apache.dubbo.rpc.protocol.AbstractInvoker.invoke(AbstractInvoker.java:187) ~[dubbo-3.2.7.jar:3.2.7]
... 295 common frames omitted
Caused by: java.lang.IllegalStateException: Not found class com.***.***DTO, cause: com.***.***DTO
at org.apache.dubbo.common.utils.ReflectUtils.forName(ReflectUtils.java:685) ~[dubbo-3.2.7.jar:3.2.7]
at org.apache.dubbo.rpc.support.RpcUtils.getParameterTypes(RpcUtils.java:169) ~[dubbo-3.2.7.jar:3.2.7]
at org.apache.dubbo.rpc.protocol.tri.TripleInvoker.invokeUnary(TripleInvoker.java:251) ~[dubbo-3.2.7.jar:3.2.7]
at org.apache.dubbo.rpc.protocol.tri.TripleInvoker.doInvoke(TripleInvoker.java:150) ~[dubbo-3.2.7.jar:3.2.7]
... 297 common frames omitted
Caused by: java.lang.ClassNotFoundException: com.***.***DTO
at org.springframework.boot.web.embedded.tomcat.TomcatEmbeddedWebappClassLoader.loadClass(TomcatEmbeddedWebappClassLoader.java:72) ~[spring-boot-2.5.12.jar:2.5.12]
at org.apache.catalina.loader.WebappClassLoaderBase.loadClass(WebappClassLoaderBase.java:1215) ~[tomcat-embed-core-9.0.60.jar:9.0.60]
at java.lang.Class.forName0(Native Method) ~[na:1.8.0_152]
at java.lang.Class.forName(Class.java:348) ~[na:1.8.0_152]
at org.apache.dubbo.common.utils.ReflectUtils.name2class(ReflectUtils.java:786) ~[dubbo-3.2.7.jar:3.2.7]
at org.apache.dubbo.common.utils.ReflectUtils.name2class(ReflectUtils.java:706) ~[dubbo-3.2.7.jar:3.2.7]
at org.apache.dubbo.common.utils.ReflectUtils.forName(ReflectUtils.java:683) ~[dubbo-3.2.7.jar:3.2.7]
... 300 common frames omitted
```
### Cause
`org.apache.dubbo.rpc.protocol.tri.TripleInvoker#invokeUnary`
When method descriptor is generic, it tries to find remote-defined-only class in client side, which causes the `ClassNotFoundException`.
```java
if (methodDescriptor instanceof StubMethodDescriptor) {
pureArgument = invocation.getArguments()[0];
} else {
if (methodDescriptor.isGeneric()) {
Object[] args = new Object[3];
args[0] = RpcUtils.getMethodName(invocation);
// here
args[1] = Arrays.stream(RpcUtils.getParameterTypes(invocation)).map(Class::getName).toArray(String[]::new);
args[2] = RpcUtils.getArguments(invocation);
pureArgument = args;
} else {
pureArgument = invocation.getArguments();
}
}
```
| https://github.com/apache/dubbo/issues/13436 | https://github.com/apache/dubbo/pull/13442 | 841ddb6e1f82d01f41a2c2a009de15a818c472be | d552cfca51a07765b080a84b8bddd044a694bc98 | 2023-11-29T09:28:06Z | java | 2023-12-01T02:25:03Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,363 | ["dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/ReflectionPackableMethod.java", "dubbo-rpc/dubbo-rpc-triple/src/test/java/org/apache/dubbo/rpc/protocol/tri/DataWrapper.java", "dubbo-rpc/dubbo-rpc-triple/src/test/java/org/apache/dubbo/rpc/protocol/tri/DescriptorService.java", "dubbo-rpc/dubbo-rpc-triple/src/test/java/org/apache/dubbo/rpc/protocol/tri/ReflectionPackableMethodTest.java"] | INTERNAL exception occurred when sending generic parameter in streaming | * Dubbo version: 3.2.5
* Operating System version: windows11
* Java version: 17
If the streaming communication mode StreamObserver generic parameter is also generic, an error "INTERNAL: Call aborted cause client exception" will be reported during runtime.
---
streaming通信模式StreamObserver泛型参数如果也是泛型的话在运行的时候会报错“INTERNAL : Call aborted cause client exception”
具体看代码截图
| https://github.com/apache/dubbo/issues/13363 | https://github.com/apache/dubbo/pull/13446 | 0d879666b4fc6163f42fcec2838633bb9f99a482 | 5554706256c9daf550f36858e5bc3d17f3ec6143 | 2023-11-15T02:56:19Z | java | 2023-12-06T02:57:09Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,222 | ["dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/ReflectionPackableMethod.java", "dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/TripleInvoker.java", "dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/TripleProtocol.java", "dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/call/ReflectionAbstractServerCall.java", "dubbo-rpc/dubbo-rpc-triple/src/test/java/org/apache/dubbo/rpc/protocol/tri/call/ReflectionServerCallTest.java"] | PackableMethod share while services cause serialization error | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Steps to reproduce this issue
In `org.apache.dubbo.rpc.protocol.tri.ReflectionPackableMethod#init`:
```java
public static ReflectionPackableMethod init(MethodDescriptor methodDescriptor, URL url) {
Object stored = methodDescriptor.getAttribute(METHOD_ATTR_PACK);
if (stored != null) {
return (ReflectionPackableMethod) stored;
}
final String serializeName = UrlUtils.serializationOrDefault(url);
final Collection<String> allSerialize = UrlUtils.allSerializations(url);
ReflectionPackableMethod reflectionPackableMethod = new ReflectionPackableMethod(
methodDescriptor, url, serializeName, allSerialize);
methodDescriptor.addAttribute(METHOD_ATTR_PACK, reflectionPackableMethod);
return reflectionPackableMethod;
}
```
Triple will cache `ReflectionPackableMethod` in `MethodDescriptor`.
But `MethodDescriptor` is shared while different services.
Thus will cause:
1. A consumer expect to use X serialization and B consumer expect to use Y serialization
2. A consumer init first, and write A serialziation into `MethodDescriptor`
3. B consumer will directly use A serialziation dur to cache
### Proper solution
Cache `ReflectionPackableMethod` in `TripleInvoker` instead of `MethodDescriptor`.
### Note
It is very easy to occur in the scenario of generic invoke. | https://github.com/apache/dubbo/issues/13222 | https://github.com/apache/dubbo/pull/13224 | 02e486a1dccef10a03b81575fb3b7158d3cd9299 | f995ef506c2e3306f9cbcd4bc33ad169a3da0b7e | 2023-10-17T09:49:21Z | java | 2023-10-18T03:39:33Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,142 | ["dubbo-compiler/src/main/resources/Dubbo3InterfaceStub.mustache", "dubbo-compiler/src/main/resources/Dubbo3TripleInterfaceStub.mustache", "dubbo-compiler/src/main/resources/Dubbo3TripleStub.mustache"] | 使用triple协议，提供方provider方法重载，consumer方消费报错，provider方空指针 | ### Environment
* Dubbo version: 3.2
* Operating System version: windows 10
* Java version:1.8
### Steps to reproduce this issue
1. 在provider侧提供重载方法如下
2. 在consumer侧分别调用两个重载方法
3. provider报错NPE
### Expected Behavior
期望能正常调用
### Actual Behavior
consumer
provider
```
Just put your stack trace here!
```
| https://github.com/apache/dubbo/issues/13142 | https://github.com/apache/dubbo/pull/13385 | 883195aeb9c92fb7e7566846e8403906f9ae027e | 4d5d9d8afc5d6d05010b8915ec72bdf9418ef02b | 2023-09-28T03:33:22Z | java | 2023-11-23T04:49:16Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,135 | ["dubbo-cluster/pom.xml", "dubbo-metrics/dubbo-metrics-api/src/main/java/org/apache/dubbo/metrics/model/MetricsSupport.java"] | dubbo 3.2 branch source code dubbo-cluster build fail | ### Environment
* Dubbo version: 3.2
* Operating System version: mac 10.15
* Java version: JDK8 or JDK 15 or JDK 17
### Steps to reproduce this issue
1. mvn clean install
2. show build error
dubbo-cluster build fail
### Expected Behavior
<!-- What do you expect from the above steps？-->
### Actual Behavior
<!-- What actually happens? -->
If there is an exception, please attach the exception trace:
error log
```
Please refer to /Users/lucas/github/dubbo/dubbo-cluster/target/surefire-reports for the individual test results.
Please refer to dump files (if any exist) [date].dump, [date]-jvmRun[N].dump and [date].dumpstream.�
```
�
| https://github.com/apache/dubbo/issues/13135 | https://github.com/apache/dubbo/pull/13139 | c4120feb8fdeb7005168b01b694417027ee84584 | 47821429807654b14010a672d20eca3d00243ffe | 2023-09-27T02:09:38Z | java | 2023-10-13T09:43:30Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,133 | ["dubbo-config/dubbo-config-api/src/main/java/org/apache/dubbo/config/bootstrap/builders/InternalServiceConfigBuilder.java"] | MetadataService should not be registered to registry | 
Step to reproduce:
- Configure `-Ddubbo.registry.address=xxx` | https://github.com/apache/dubbo/issues/13133 | https://github.com/apache/dubbo/pull/13134 | bd5d158d1bb4295837b4b79e08f6ae5230b315c1 | 414608245329fbafd00a68dd1d53a0cd97ad3b6f | 2023-09-26T09:21:01Z | java | 2023-09-26T10:55:08Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,078 | ["dubbo-remoting/dubbo-remoting-api/src/main/java/org/apache/dubbo/remoting/Constants.java", "dubbo-remoting/dubbo-remoting-netty4/src/main/java/org/apache/dubbo/remoting/transport/netty4/NettyConnectionClient.java", "dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/TripleHttp2Protocol.java", "dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/transport/TripleGoAwayHandler.java"] | tri协议Http2NoMoreStreamIdsException异常 | * Dubbo version: 3.2.4
use tri protocol， it has this exception when service run many days，it can not resolve itself，we must restart the service
基本跑几天会出来，然后一直报错无法恢复，目前只能自动重启或者手动重启重连
the problem maybe is not reuse the （unary mode） stream
初步怀疑是unary的stream复用问题
| https://github.com/apache/dubbo/issues/13078 | https://github.com/apache/dubbo/pull/13228 | 3f07e683a184670ebfc0ef48c268836931b39a75 | 02e486a1dccef10a03b81575fb3b7158d3cd9299 | 2023-09-19T03:12:25Z | java | 2023-10-18T03:38:57Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,042 | ["dubbo-common/src/main/java/org/apache/dubbo/common/threadpool/support/AbortPolicyWithReport.java"] | AbortPolicyWithReport 高并发情况下 dump 多次的问题 | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 2.7.x, 3.x
* Operating System version: ALL
* Java version: ALL
### Steps to reproduce this issue
高并发请求时, 偶发复现
### Expected Behavior
间隔指定时间, 比如 10 分钟, 再进行 threaddump
### Actual Behavior
下面是服务器上的日志记录, 可以看到 dubbo 线程池耗尽时的 threaddump 并没有按照预期的间隔 10分钟进行
```
-rw-rw-r-- 1 root root 695929 Sep 1 19:30 Dubbo_JStack.log.2023-09-01_19:30:39
-rw-rw-r-- 1 root root 1068779 Sep 3 09:00 Dubbo_JStack.log.2023-09-03_09:00:35
-rw-rw-r-- 1 root root 2326331 Sep 3 23:53 Dubbo_JStack.log.2023-09-03_23:53:21
-rw-rw-r-- 1 root root 1081798 Sep 5 09:00 Dubbo_JStack.log.2023-09-05_09:00:35
-rw-rw-r-- 1 root root 1652182 Sep 5 09:00 Dubbo_JStack.log.2023-09-05_09:00:38
-rw-rw-r-- 1 root root 956693 Sep 9 19:30 Dubbo_JStack.log.2023-09-09_19:30:34
-rw-rw-r-- 1 root root 1515179 Sep 9 19:30 Dubbo_JStack.log.2023-09-09_19:30:37
-rw-rw-r-- 1 root root 981556 Sep 10 09:00 Dubbo_JStack.log.2023-09-10_09:00:38
-rw-rw-r-- 1 root root 1443054 Sep 10 09:00 Dubbo_JStack.log.2023-09-10_09:00:42
-rw-rw-r-- 1 root root 999105 Sep 11 09:00 Dubbo_JStack.log.2023-09-11_09:00:38
-rw-rw-r-- 1 root root 1500287 Sep 11 09:00 Dubbo_JStack.log.2023-09-11_09:00:43
```
### 原因分析
按照以下顺序获取锁就可能导致前面的现象, 线程切换 + 错误的顺序设置 lastPrintTime 共同导致了这个问题
```
thread1: 1. 检查10分钟间隔通过 3. 获取锁 4. threaddump(stw) 5. 释放锁 7. 设置 lastPrintTime
thread2: 2. 检查10分钟间隔通过 6. 获取锁 8. threaddump(stw) 9. 释放锁
```
源码如下 https://github.com/giraffe-tree/dubbo/blob/3.2/dubbo-common/src/main/java/org/apache/dubbo/common/threadpool/support/AbortPolicyWithReport.java
```java
//dump every 10 minutes
if (now - lastPrintTime < TEN_MINUTES_MILLS) {
return;
}
if (!guard.tryAcquire()) {
return;
}
ExecutorService pool = Executors.newSingleThreadExecutor();
pool.execute(() -> {
// ....代码省略
try (FileOutputStream jStackStream = new FileOutputStream(
new File(dumpPath, "Dubbo_JStack.log" + "." + dateStr))) {
jstack(jStackStream);
} catch (Exception t) {
logger.error(COMMON_UNEXPECTED_CREATE_DUMP, "", "", "dump jStack error", t);
} finally {
guard.release();
}
lastPrintTime = System.currentTimeMillis();
});
//must shutdown thread pool ,if not will lead to OOM
pool.shutdown();
```
### 修复方案
1. 在获取锁后再次判断是否满足时间要求
2. 在释放锁前设置 lastPrintTime
```java
if (now - lastPrintTime < TEN_MINUTES_MILLS) {
return;
}
if (!guard.tryAcquire()) {
return;
}
// 1. 此处再次判断时间
if (System.currentTimeMillis() - lastPrintTime < TEN_MINUTES_MILLS) {
return;
}
ExecutorService pool = Executors.newSingleThreadExecutor();
pool.execute(() -> {
// ....代码省略
try (FileOutputStream jStackStream = new FileOutputStream(
new File(dumpPath, "Dubbo_JStack.log" + "." + dateStr))) {
jstack(jStackStream);
} catch (Exception t) {
logger.error(COMMON_UNEXPECTED_CREATE_DUMP, "", "", "dump jStack error", t);
} finally {
// 2. 在释放锁前设置 lastPrintTime
lastPrintTime = System.currentTimeMillis();
guard.release();
}
});
//must shutdown thread pool ,if not will lead to OOM
pool.shutdown();
```
| https://github.com/apache/dubbo/issues/13042 | https://github.com/apache/dubbo/pull/13043 | e887bab83b3582351a92b8a893610a46545517b8 | d9db031469c27eb2e1c7438cf92e217b7f6817c4 | 2023-09-11T15:07:48Z | java | 2023-09-13T01:54:48Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 13,030 | ["dubbo-filter/dubbo-filter-validation/src/main/java/org/apache/dubbo/validation/filter/ValidationFilter.java", "dubbo-filter/dubbo-filter-validation/src/main/java/org/apache/dubbo/validation/support/jvalidation/JValidator.java", "dubbo-filter/dubbo-filter-validation/src/main/java/org/apache/dubbo/validation/support/jvalidation/JValidatorNew.java", "dubbo-filter/dubbo-filter-validation/src/test/java/org/apache/dubbo/validation/support/jvalidation/JValidatorTest.java", "dubbo-filter/dubbo-filter-validation/src/test/java/org/apache/dubbo/validation/support/jvalidation/mock/JValidatorTestTarget.java"] | Support ValidationFilter to display parameter names when validation fails | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
- [x] I have searched the [release notes](https://github.com/apache/dubbo/releases) of this repository and believe that this is not a duplicate.
## Describe the feature
<!-- Please also discuss possible business value --> | https://github.com/apache/dubbo/issues/13030 | https://github.com/apache/dubbo/pull/13029 | d9db031469c27eb2e1c7438cf92e217b7f6817c4 | 8a509e9601fab4af278859c9d947fca8f5f0fa27 | 2023-09-08T19:16:08Z | java | 2023-09-13T09:12:25Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,945 | ["dubbo-remoting/dubbo-remoting-netty4/src/main/java/org/apache/dubbo/remoting/transport/netty4/NettyConnectionClient.java", "dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/TripleHttp2Protocol.java", "dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/TriplePingPongHandler.java"] | 服务端突然断网、断电，客户端无法进行故障转移 | dubbo 版本 3.2.4
实验过程：服务端关闭网络，节点在15min内不会从nacos平台消失。导致consumer仍然能够拿到invoder，此时客户端如果能判断服务端invoker不能连接，就会从全局的invoker列表中剔除该invoker，执行故障转移。但是dubbo consumer 在服务端突然断网后，判断连接仍然存在，导致无法剔除invoker，无法进行故障转移。
问题：针对服务端突然断电、断网，是否是应为tcp链接无法正常断开，导致consumer判断连接失效？
<img width="916" alt="image" src="https://github.com/apache/dubbo/assets/139483558/38486a89-695a-4792-83a6-51f783f5abf2">
| https://github.com/apache/dubbo/issues/12945 | https://github.com/apache/dubbo/pull/12955 | a4ca07087e0bd4c3358ec7f095b868cc707dda32 | b84c1386bba2ac306a002b1a283cdb952b456047 | 2023-08-22T10:09:33Z | java | 2023-10-11T03:17:41Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,927 | ["dubbo-rpc/dubbo-rpc-injvm/src/main/java/org/apache/dubbo/rpc/protocol/injvm/InjvmInvoker.java"] | 使用手脚架生成项目启动包错 | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.2.4
* Operating System version: windows 11
* Java version: 17.0.7
使用dubbo 手脚架生成的项目没有改任何代码启动报错
java.lang.reflect.InaccessibleObjectException: Unable to make field final int java.math.BigInteger.signum accessible: module java.base does not "opens java.math" to unnamed module @73a1e9a9
at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:354)
at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:297)
at java.base/java.lang.reflect.Field.checkCanSetAccessible(Field.java:178)
at java.base/java.lang.reflect.Field.setAccessible(Field.java:172)
at com.alibaba.com.caucho.hessian.io.JavaDeserializer.getFieldMap(JavaDeserializer.java:340)
at com.alibaba.com.caucho.hessian.io.JavaDeserializer.<init>(JavaDeserializer.java:80)
at com.alibaba.com.caucho.hessian.io.BigIntegerDeserializer.<init>(BigIntegerDeserializer.java:25)
at com.alibaba.com.caucho.hessian.io.SerializerFactory.<clinit>(SerializerFactory.java:142)
at org.apache.dubbo.common.serialize.hessian2.Hessian2FactoryManager.createDefaultSerializerFactory(Hessian2FactoryManager.java:81)
at org.apache.dubbo.common.serialize.hessian2.Hessian2FactoryManager.createSerializerFactory(Hessian2FactoryManager.java:77)
at org.apache.dubbo.common.serialize.hessian2.Hessian2FactoryManager.getSerializerFactory(Hessian2FactoryManager.java:62)
at org.apache.dubbo.common.serialize.hessian2.Hessian2ObjectOutput.<init>(Hessian2ObjectOutput.java:44)
at org.apache.dubbo.common.serialize.hessian2.Hessian2Serialization.serialize(Hessian2Serialization.java:57)
at org.apache.dubbo.common.serialize.DefaultSerializationExceptionWrapper.serialize(DefaultSerializationExceptionWrapper.java:51)
at org.apache.dubbo.rpc.protocol.injvm.DefaultParamDeepCopyUtil.copy(DefaultParamDeepCopyUtil.java:45)
at org.apache.dubbo.rpc.protocol.injvm.InjvmInvoker.recreateInvocation(InjvmInvoker.java:245)
at org.apache.dubbo.rpc.protocol.injvm.InjvmInvoker.doInvoke(InjvmInvoker.java:119)
at org.apache.dubbo.rpc.protocol.AbstractInvoker.doInvokeAndReturn(AbstractInvoker.java:243)
at org.apache.dubbo.rpc.protocol.AbstractInvoker.invoke(AbstractInvoker.java:187)
at org.apache.dubbo.rpc.listener.ListenerInvokerWrapper.invoke(ListenerInvokerWrapper.java:71)
at org.apache.dubbo.metrics.filter.MetricsFilter.invoke(MetricsFilter.java:63)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:334)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.invoke(FilterChainBuilder.java:196)
at org.apache.dubbo.rpc.protocol.ReferenceCountInvokerWrapper.invoke(ReferenceCountInvokerWrapper.java:78)
at org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invokeWithContext(AbstractClusterInvoker.java:383)
at org.apache.dubbo.rpc.cluster.support.FailoverClusterInvoker.doInvoke(FailoverClusterInvoker.java:80)
at org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invoke(AbstractClusterInvoker.java:344)
at org.apache.dubbo.rpc.cluster.router.RouterSnapshotFilter.invoke(RouterSnapshotFilter.java:46)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:334)
at org.apache.dubbo.monitor.support.MonitorFilter.invoke(MonitorFilter.java:108)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:334)
at org.apache.dubbo.rpc.cluster.filter.support.MetricsClusterFilter.invoke(MetricsClusterFilter.java:50)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:334)
at org.apache.dubbo.rpc.protocol.dubbo.filter.FutureFilter.invoke(FutureFilter.java:52)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:334)
at org.apache.dubbo.spring.security.filter.ContextHolderParametersSelectedTransferFilter.invoke(ContextHolderParametersSelectedTransferFilter.java:41)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:334)
at org.apache.dubbo.rpc.cluster.filter.support.ConsumerClassLoaderFilter.invoke(ConsumerClassLoaderFilter.java:40)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:334)
at org.apache.dubbo.rpc.cluster.filter.support.ConsumerContextFilter.invoke(ConsumerContextFilter.java:118)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:334)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.invoke(FilterChainBuilder.java:196)
at org.apache.dubbo.rpc.cluster.support.wrapper.AbstractCluster$ClusterFilterInvoker.invoke(AbstractCluster.java:91)
at org.apache.dubbo.rpc.cluster.support.wrapper.MockClusterInvoker.invoke(MockClusterInvoker.java:103)
at org.apache.dubbo.rpc.cluster.support.wrapper.ScopeClusterInvoker.invoke(ScopeClusterInvoker.java:163)
at org.apache.dubbo.rpc.proxy.InvocationUtil.invoke(InvocationUtil.java:57)
at org.apache.dubbo.rpc.proxy.InvokerInvocationHandler.invoke(InvokerInvocationHandler.java:75)
at cn.charlab.dubbotest.dubbo.api.DemoServiceDubboProxy0.sayHello(DemoServiceDubboProxy0.java)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77)
at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.base/java.lang.reflect.Method.invoke(Method.java:568)
at org.apache.dubbo.config.spring.util.LazyTargetInvocationHandler.invoke(LazyTargetInvocationHandler.java:58)
at cn.charlab.dubbotest.dubbo.api.DemoServiceDubboProxy0.sayHello(DemoServiceDubboProxy0.java)
at cn.charlab.dubbotest.dubbo.consumer.Consumer.run(Consumer.java:18)
at org.springframework.boot.SpringApplication.callRunner(SpringApplication.java:769)
at org.springframework.boot.SpringApplication.callRunners(SpringApplication.java:753)
at org.springframework.boot.SpringApplication.run(SpringApplication.java:317)
at org.springframework.boot.SpringApplication.run(SpringApplication.java:1304)
at org.springframework.boot.SpringApplication.run(SpringApplication.java:1293)
at cn.charlab.dubbotest.DubbotestApplication.main(DubbotestApplication.java:12)
| https://github.com/apache/dubbo/issues/12927 | https://github.com/apache/dubbo/pull/12938 | 0c6b1ee17ada37b1363421f29cc6275dc27d716e | c2352df095c86b50496eea7d31d569fe02f8ebac | 2023-08-21T05:56:58Z | java | 2023-08-23T02:19:10Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,916 | ["dubbo-remoting/dubbo-remoting-http/src/main/java/org/apache/dubbo/remoting/http/restclient/HttpClientRestClient.java"] | 3.2.5 rest协议使用apache-http-client出现NPE | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.2.5
* Operating System version: MAC OS
* Java version: 17
### Steps to reproduce this issue
1. 定义rest接口并配置client=apache-http-client
```
**
* 文件管理服务接口
*
* @author chaoxin.lu
* @since 2018/6/28
*/
@RequestMapping("/oss")
public interface FileSvc {
/**
* 获取文件
*
* @param bucket
* @param key
* @return
*/
@GetMapping
@ResponseBody
R<FileEntry> getObject(@RequestParam("bucket") String bucket,
@RequestParam("key") String key);
}
```
```
dubbo:
protocol:
id: ${DUBBO_PROTOCOL:rest}
name: ${DUBBO_PROTOCOL:rest}
client: ${DUBBO_CLIENT:apache-http-client}
port: ${DUBBO_PORT:20895}
server: ${DUBBO_PROTOCOL_SERVER:netty}
```
3. 服务消费者如下
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
<!-- What do you expect from the above steps？-->
### Actual Behavior
<!-- What actually happens? -->
If there is an exception, please attach the exception trace:
```
Caused by: java.lang.NullPointerException: Cannot invoke "org.apache.http.Header.getValue()" because the return value of "org.apache.http.client.methods.CloseableHttpResponse.getFirstHeader(String)" is null
at org.apache.dubbo.remoting.http.restclient.HttpClientRestClient$1.getContentType(HttpClientRestClient.java:96)
at org.apache.dubbo.rpc.protocol.rest.RestInvoker.lambda$doInvoke$0(RestInvoker.java:105)
at java.base/java.util.concurrent.CompletableFuture.uniWhenComplete(CompletableFuture.java:863)
at java.base/java.util.concurrent.CompletableFuture.uniWhenCompleteStage(CompletableFuture.java:887)
at java.base/java.util.concurrent.CompletableFuture.whenComplete(CompletableFuture.java:2325)
at org.apache.dubbo.rpc.protocol.rest.RestInvoker.doInvoke(RestInvoker.java:87)
at org.apache.dubbo.rpc.protocol.AbstractInvoker.doInvokeAndReturn(AbstractInvoker.java:243)
at org.apache.dubbo.rpc.protocol.AbstractInvoker.invoke(AbstractInvoker.java:187)
... 97 more
```
NPE点如下
使用默认的okhttp则没有该问题
| https://github.com/apache/dubbo/issues/12916 | https://github.com/apache/dubbo/pull/12984 | cdd7a405641c994a55d85690b87291867bee6820 | 35746aaf5a3e6eab9181fedb37fd8cc00793469c | 2023-08-16T11:20:26Z | java | 2023-09-05T03:22:21Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,779 | ["dubbo-spring-boot/dubbo-spring-boot-compatible/autoconfigure/src/main/java/org/apache/dubbo/spring/boot/env/DubboDefaultPropertiesEnvironmentPostProcessor.java", "dubbo-spring-boot/dubbo-spring-boot-compatible/autoconfigure/src/test/java/org/apache/dubbo/spring/boot/env/DubboDefaultPropertiesEnvironmentPostProcessorTest.java"] | 通过外部配置文件的“dubbo.application.qos-enable=false”配置来关闭QOS不生效 | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.2.4
* Operating System version: win10
* Java version: 17
### Steps to reproduce this issue
1. 启动dubbo服务时加入JVM参数：-Ddubbo.properties.file=/data/dubbo/dubbo.properties
2. /data/dubbo/dubbo.properties 中配置：dubbo.application.qos-enable=false
3. 检查 QOS端口是否打开（已打开）
4. 通过监控地址：http://localhost:9002/actuator/dubbo/properties
查看 qos-enable 属性（true）
```
{
"dubbo.application.name": "hp-demo-infra-service",
"dubbo.application.organization": "HP-Socket",
"dubbo.application.owner": "Kingfisher",
"dubbo.application.qos-enable": "true",
"dubbo.application.qos-port": "7002",
"dubbo.application.version": "0.0.1-SNAPSHOT",
"dubbo.config.multiple": "true",
"dubbo.properties.file": "/data/dubbo/dubbo.properties",
"dubbo.protocol.name": "dubbo",
"dubbo.protocol.port": "6002",
"dubbo.registry.file": "/data/dubbo/.cache/dubbo-registry_hp-demo-infra-service.properties",
"dubbo.resolve.file": "/data/dubbo/dubbo-resolve.properties",
"dubbo.scan.base-packages": "com.github.hpsocket.demo"
}
```
而通过JVM参数：“-Ddubbo.application.qos-enable=false”，或在 application.yml 中定义 ”dubbo.application.qos-enable: false”，配置是生效的。
| https://github.com/apache/dubbo/issues/12779 | https://github.com/apache/dubbo/pull/12861 | b87746b19fb99b60b3e6973b3b50e3daca04eb27 | 7290ae59bc69577b5b902dc0f6e161eb4c74db6b | 2023-07-24T02:38:58Z | java | 2023-08-16T13:01:28Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,773 | ["dubbo-common/src/main/java/org/apache/dubbo/common/constants/CommonConstants.java", "dubbo-monitor/dubbo-monitor-api/src/main/java/org/apache/dubbo/monitor/support/CallbackConsumerContextFilter.java", "dubbo-monitor/dubbo-monitor-api/src/main/resources/META-INF/dubbo/internal/org.apache.dubbo.rpc.Filter", "dubbo-rpc/dubbo-rpc-dubbo/src/main/java/org/apache/dubbo/rpc/protocol/dubbo/CallbackServiceCodec.java"] | Callback mode rpc context loss, which also affects link tracking | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.1.x
* Operating System version: mac os\Linux
* Java version: 1.8
### Steps to reproduce this issue
1. Server
```java
public interface DemoApi {
String methodA();
String testCallbackMethod(String name, ApiCallback callback);
interface ApiCallback {
String callbackMethod();
}
}
@DubboService(methods = @Method(name="testCallbackMethod",
arguments = {@Argument(index = 1, type = "DemoApi$ApiCallback", callback = true)}))
public class DemoProvider implements DemoApi {
@Override
public String methodA() {
return "ok";
}
@Override
public String testCallbackMethod(String name, ApiCallback callback) {
String s = callback.callbackMethod();
return "ok";
}
}
```
2. Consumer
```java
@RestController
public class TestController{
@DubboReference
DemoApi demo;
private DemoApi.ApiCallback callback = new DemoApi.ApiCallback() {
@Override
public String callbackMethod() {
//rpc context has been lost
System.out.println("callback:" + RpcContext.getContext().getAttachment("test"));
//This is just a simple simulation problem, in fact, it may call other services
return demo.methodA();
}
};
@GetMapping("/testDubboCallback")
public ResponseEntity<String> testDubboCallback() {
RpcContext.getContext().setAttachment("test", "abc");
return ResponseEntity.ok(demo.testCallbackMethod("param" ,callback));
}
}
```
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
<!-- What do you expect from the above steps？-->
RpcContext passed normally， sky tracking is normal.
### Actual Behavior
<!-- What actually happens? -->
rpc context loss, the tracking has been broken.
If there is an exception, please attach the exception trace:
```
Just put your stack trace here!
```
| https://github.com/apache/dubbo/issues/12773 | https://github.com/apache/dubbo/pull/12866 | bac06929f21e42ba8fc574f3e829341432d571e2 | 78a4dbad6e5ed195acebc4d2eb18588c79563367 | 2023-07-21T07:04:20Z | java | 2023-08-09T08:05:27Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,765 | ["dubbo-config/dubbo-config-spring/src/main/java/org/apache/dubbo/config/spring/beans/factory/annotation/ServiceAnnotationPostProcessor.java"] | 关于 @Service 注解日志警告的问题 | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
## Ask your question here
现在都用 @DubboService 注解了 然而每次启动还会报 @Service 扫描不到的警告 这是不是有点非必要啊
| https://github.com/apache/dubbo/issues/12765 | https://github.com/apache/dubbo/pull/12981 | 8595711e99f9102eb1055a80a4698e118c82cfb7 | cdd7a405641c994a55d85690b87291867bee6820 | 2023-07-20T09:22:04Z | java | 2023-09-05T02:50:35Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,698 | ["dubbo-dependencies-bom/pom.xml"] | dubbo3.0在nacos配置中心修改消费者的订阅模式，是实时生效的吗？ | dubbo3.0在nacos配置中心修改消费者的订阅模式，是实时生效的吗？比如前期是双订阅模式，然后把它改成仅应用级注册后，需要重启服务才能生效吗 | https://github.com/apache/dubbo/issues/12698 | https://github.com/apache/dubbo/pull/12304 | ce6e5a6eedd9c8916e90d31d2fe5b732789f546b | 53255e3f8e0cce46c0e6f9e5261ad24eb9d38509 | 2023-07-08T10:37:32Z | java | 2023-05-16T04:53:40Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,658 | ["dubbo-remoting/dubbo-remoting-api/src/main/java/org/apache/dubbo/remoting/api/connection/AbstractConnectionClient.java", "dubbo-remoting/dubbo-remoting-netty4/src/main/java/org/apache/dubbo/remoting/transport/netty4/NettyConnectionClient.java"] | dubbo provider cannot reconnect service without registry center since 3.2.1 | Description:
a dubbo provider project with springboot invokes dubbo service, defining referenced service like this:
```
@DubboReference( check = false,url = "tri://localhost:50051")
private HelloService helloService;
```
The consumer repeatedly calls the service with a delay of one second after each call.
```
@Scheduled(fixedDelay = 1000)
public void run() {
String message = helloService.sayHello("lucky");
System.out.println(message);
}
```
key config items in application.properties like this :
```
dubbo.registry.address: N/A
```
Question:
The above code can run normally in dubbo-3.1.7, when starting the service consumer first, and then starting the service provider. At 3.2.1, the service consumer could not successfully reconnect to the service provider.
According to the source code, the implementation of the ConnectionListener class, which is responsible for reconnecting, has changed. The number of connection references in the code changed from positive to 0, causing the logic to terminate early. Why is that？
Partial Code of Reconnection:
| https://github.com/apache/dubbo/issues/12658 | https://github.com/apache/dubbo/pull/12737 | bd76d044ba7118409a4bb52c1fa2a0f2cbc9e7a3 | 630e14ae404d71dbd7229e3205468b3069b60676 | 2023-07-03T11:35:31Z | java | 2023-07-17T03:35:03Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,650 | ["dubbo-common/src/main/java/org/apache/dubbo/common/logger/log4j2/Log4j2LoggerAdapter.java", "dubbo-common/src/main/java/org/apache/dubbo/common/logger/slf4j/Slf4jLogger.java", "dubbo-common/src/main/java/org/apache/dubbo/common/logger/slf4j/Slf4jLoggerAdapter.java"] | 应用初始化启动后，使用telnet查看“日志框架运行时”抛NPE | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.1.9
* Operating System version: Windows
* Java version: 11.0.18
### Steps to reproduce this issue
1. 正常启动一个普通springboot+dubbo应用，使用默认的logback且未配置日志选项
2. 使用 QoS 命令 `loggerInfo` 查询日志配置
3. 未正确返回结果，且应用控制台抛出NPE
```bash
telnet localhost 22222
___ __ __ ___ ___ ____
/ _ \ / / / // _ ) / _ ) / __ \
/ // // /_/ // _ |/ _ |/ /_/ /
/____/ \____//____//____/ \____/
dubbo>loggerInfo
loggerInfo :fail to execute commandContext by null
```
### Expected Behavior
<!-- What do you expect from the above steps？-->
正常查询应用的日志配置 ，如果某些参数未配置则提供默认值（如 Log level: INFO）
### Actual Behavior
<!-- What actually happens? -->
If there is an exception, please attach the exception trace:
```
14:15:36.417 logback [qos-worker-3-1] INFO o.a.d.q.c.DefaultCommandExecutor - [DUBBO] [Dubbo QoS] Command Process start. Command: loggerInfo, Args: [], Remote Address: /0:0:0:0:0:0:0:1:63671, dubbo version: 3.1.9, current host: 10.201.50.35
No such logging.properties in classpath for jdk logging config!
14:15:36.450 logback [qos-worker-3-1] INFO o.a.d.q.c.DefaultCommandExecutor - [DUBBO] [Dubbo QoS] Command Process Failed. Command: loggerInfo, Args: [], Remote Address: /0:0:0:0:0:0:0:1:63671, dubbo version: 3.1.9, current host: 10.201.50.35
java.lang.NullPointerException: null
at org.apache.dubbo.qos.command.impl.LoggerInfo.execute(LoggerInfo.java:35)
at org.apache.dubbo.qos.command.DefaultCommandExecutor.execute(DefaultCommandExecutor.java:57)
at org.apache.dubbo.qos.server.handler.TelnetProcessHandler.channelRead0(TelnetProcessHandler.java:59)
at org.apache.dubbo.qos.server.handler.TelnetProcessHandler.channelRead0(TelnetProcessHandler.java:40)
14:15:36.451 logback [qos-worker-3-1] ERROR o.a.d.q.s.h.TelnetProcessHandler - [DUBBO] execute commandContext got exception org.apache.dubbo.qos.command.CommandContext@740b26f8, dubbo version: 3.1.9, current host: 10.201.50.35, error code: 7-6. This may be caused by , go to https://dubbo.apache.org/faq/7/6 to find instructions.
java.lang.NullPointerException: null
at org.apache.dubbo.qos.command.impl.LoggerInfo.execute(LoggerInfo.java:35)
at org.apache.dubbo.qos.command.DefaultCommandExecutor.execute(DefaultCommandExecutor.java:57)
at org.apache.dubbo.qos.server.handler.TelnetProcessHandler.channelRead0(TelnetProcessHandler.java:59)
at org.apache.dubbo.qos.server.handler.TelnetProcessHandler.channelRead0(TelnetProcessHandler.java:40)
```
断点调试发现 `org.apache.dubbo.qos.command.impl.LoggerInfo` 33行中返回了 null
使用 `switchLogLevel` 修改日志级别后再使用 `loggerInfo` 查询则正常
```
dubbo>loggerInfo
loggerInfo :fail to execute commandContext by null
dubbo>switchLogLevel INFO
OK
dubbo>loggerInfo
Available logger adapters: [slf4j, jcl, log4j2, jdk]. Current Adapter: [slf4j]. Log level: INFO
``` | https://github.com/apache/dubbo/issues/12650 | https://github.com/apache/dubbo/pull/12671 | 7c7f3db772589932bd58534fba8b1864d46d394b | 879fd811283c28a5fcb4dfebd383e05b5e500cb0 | 2023-07-03T06:27:11Z | java | 2023-07-25T13:18:30Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,646 | ["dubbo-registry/dubbo-registry-zookeeper/src/main/java/org/apache/dubbo/registry/zookeeper/ZookeeperRegistry.java"] | ZookeeperRegistry ->doRegister 方法 faultTolerant参数 设置为false，是不是会存在问题 ？ | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
## Ask your question here
[这个问题中](https://github.com/apache/dubbo/issues/12607)
在provider 恢复zk 链接后
日志如下
```
2023-06-27 19:24:53.473 [main-EventThread] INFO org.apache.curator.framework.state.ConnectionStateManager [postState:251] - State change: RECONNECTED
```
FailbackRegistry 尝试去恢复链接 会调用 addFailedRegistered 加入到HashedWheelTimer
HashedWheelTimer 最后尝试去重新注册的时候提示node 已经存在
```
at org.apache.zookeeper.KeeperException.create(KeeperException.java:119)
at org.apache.zookeeper.KeeperException.create(KeeperException.java:51)
at org.apache.zookeeper.ZooKeeper.create(ZooKeeper.java:783)
at org.apache.curator.framework.imps.CreateBuilderImpl$17.call(CreateBuilderImpl.java:1176)
at org.apache.curator.framework.imps.CreateBuilderImpl$17.call(CreateBuilderImpl.java:1156)
at org.apache.curator.connection.StandardConnectionHandlingPolicy.callWithRetry(StandardConnectionHandlingPolicy.java:64)
at org.apache.curator.RetryLoop.callWithRetry(RetryLoop.java:100)
at org.apache.curator.framework.imps.CreateBuilderImpl.pathInForeground(CreateBuilderImpl.java:1153)
at org.apache.curator.framework.imps.CreateBuilderImpl.protectedPathInForeground(CreateBuilderImpl.java:607)
at org.apache.curator.framework.imps.CreateBuilderImpl.forPath(CreateBuilderImpl.java:597)
at org.apache.curator.framework.imps.CreateBuilderImpl.forPath(CreateBuilderImpl.java:575)
at org.apache.curator.framework.imps.CreateBuilderImpl.forPath(CreateBuilderImpl.java:51)
at org.apache.dubbo.remoting.zookeeper.curator.CuratorZookeeperClient.createEphemeral(CuratorZookeeperClient.java:138)
at org.apache.dubbo.remoting.zookeeper.AbstractZookeeperClient.create(AbstractZookeeperClient.java:87)
at org.apache.dubbo.registry.zookeeper.ZookeeperRegistry.doRegister(ZookeeperRegistry.java:161)
at org.apache.dubbo.registry.retry.FailedRegisteredTask.doRetry(FailedRegisteredTask.java:37)
at org.apache.dubbo.registry.retry.AbstractRetryTask.run(AbstractRetryTask.java:129)
at org.apache.dubbo.common.timer.HashedWheelTimer$HashedWheelTimeout.expire(HashedWheelTimer.java:653)
at org.apache.dubbo.common.timer.HashedWheelTimer$HashedWheelBucket.expireTimeouts(HashedWheelTimer.java:732)
at org.apache.dubbo.common.timer.HashedWheelTimer$Worker.run(HashedWheelTimer.java:454)
at java.base/java.lang.Thread.run(Thread.java:833)
```
这个问题与 [瓜子遇到的问题类似](https://cn.dubbo.apache.org/zh-cn/blog/2019/01/05/dubbo-%e5%9c%a8%e7%93%9c%e5%ad%90%e4%ba%8c%e6%89%8b%e8%bd%a6%e7%9a%84%e5%ae%9e%e8%b7%b5/)
这块代码好像没有在3.0的代码中 而是用了faultTolerant 这个参数
查看了源码 AbstractZookeeperClient 的 create()只有在 ZookeeperRegistry 调用 doRegister 设置为false。
提交日志 显示#11036 这个issue后修改的doRegisterd的faultTolerant 从true 改为了 false
| https://github.com/apache/dubbo/issues/12646 | https://github.com/apache/dubbo/pull/12740 | 7a51c2379f4f7c4c6c6ed2ce53ae178304cbcc7b | c54e1b63f975d18797a85046cf23dfcc989d3456 | 2023-07-03T01:07:44Z | java | 2023-07-19T06:55:18Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,620 | ["dubbo-config/dubbo-config-spring/src/main/java/org/apache/dubbo/config/spring/context/DubboDeployApplicationListener.java"] | dubbo3.1.5版本BUG！！！，启动是出现线程死锁导致启动服务失败 | 升级dubbo3遇到问题，项目里使用@PostConstruct注解，并且方法里调用的是使用了@Scheduled注解的方法，造成线程死锁
日志信息：
```
"Thread-54" #617 daemon prio=5 os_prio=0 tid=0x00007f55ca668000 nid=0x26d waiting for monitor entry [0x00007f5554af6000]
java.lang.Thread.State: BLOCKED (on object monitor)
at org.apache.dubbo.config.spring.ReferenceBean.getCallProxy(ReferenceBean.java:351) 【【【dubbo3.1.5版本引出的问题】】】】
- waiting to lock <0x000000069092aa18> (a java.util.concurrent.ConcurrentHashMap)
at org.apache.dubbo.config.spring.ReferenceBean.access$100(ReferenceBean.java:100)
at org.apache.dubbo.config.spring.ReferenceBean$DubboReferenceLazyInitTargetSource.createObject(ReferenceBean.java:359)
at org.springframework.aop.target.AbstractLazyCreationTargetSource.getTarget(AbstractLazyCreationTargetSource.java:86)
- locked <0x00000005d7866948> (a org.apache.dubbo.config.spring.ReferenceBean$DubboReferenceLazyInitTargetSource)
at org.springframework.aop.framework.JdkDynamicAopProxy.invoke(JdkDynamicAopProxy.java:192)
at com.sun.proxy.$Proxy73.queryAllData(Unknown Source)
at com.huasheng.core.client.basicData.EtfDataServiceClient.queryAllData$original$f6OsWsrc(EtfDataServiceClient.java:41)
at com.huasheng.core.client.basicData.EtfDataServiceClient.queryAllData$original$f6OsWsrc$accessor$ggthnbjE(EtfDataServiceClient.java)
at com.huasheng.core.client.basicData.EtfDataServiceClient$auxiliary$x5YqXG7G.call(Unknown Source)
at org.apache.skywalking.apm.agent.core.plugin.interceptor.enhance.InstMethodsInter.intercept(InstMethodsInter.java:86)
at com.huasheng.core.client.basicData.EtfDataServiceClient.queryAllData(EtfDataServiceClient.java)
at com.huasheng.task.job.hk.HqBasicDataRefreshJob.refreshindexEtf$original$UrAXt9j7(HqBasicDataRefreshJob.java:1013)
at com.huasheng.task.job.hk.HqBasicDataRefreshJob.refreshindexEtf$original$UrAXt9j7$accessor$WzXLmL3m(HqBasicDataRefreshJob.java)
at com.huasheng.task.job.hk.HqBasicDataRefreshJob$auxiliary$J3bNxn2m.call(Unknown Source)
at org.apache.skywalking.apm.agent.core.plugin.interceptor.enhance.InstMethodsInter.intercept(InstMethodsInter.java:86)
at com.huasheng.task.job.hk.HqBasicDataRefreshJob.refreshindexEtf(HqBasicDataRefreshJob.java)
at com.huasheng.task.job.hk.HqBasicDataRefreshJob.lambda$indexEtfJob$4(HqBasicDataRefreshJob.java:1006)
at com.huasheng.task.job.hk.HqBasicDataRefreshJob$$Lambda$452/1550461532.run(Unknown Source)
at java.lang.Thread.run(Thread.java:748)
Locked ownable synchronizers:
- None
```
| https://github.com/apache/dubbo/issues/12620 | https://github.com/apache/dubbo/pull/12659 | 78ca5e523cd6969a0b7085e51d3996fd8448a55c | 1c5265073bb55d7a697438a2bae944377fb89343 | 2023-06-29T07:01:24Z | java | 2023-07-14T01:52:11Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,545 | ["dubbo-cluster/src/main/java/org/apache/dubbo/rpc/cluster/support/wrapper/ScopeClusterInvoker.java", "dubbo-rpc/dubbo-rpc-injvm/src/main/java/org/apache/dubbo/rpc/protocol/injvm/InjvmInvoker.java"] | dubbo 3.2.2 使用 injvm 调用 报反射异常 | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.2.2
* Operating System version: xxx
* Java version: 17
* spring cloud alibaba 2022
### Steps to reproduce this issue
```
java.lang.reflect.InaccessibleObjectException: Unable to make field private byte java.lang.StackTraceElement.format accessible: module java.base does not "opens java.lang" to unnamed module @74751b3
at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:354)
at java.base/java.lang.reflect.AccessibleObject.checkCanSetAccessible(AccessibleObject.java:297)
at java.base/java.lang.reflect.Field.checkCanSetAccessible(Field.java:178)
at java.base/java.lang.reflect.Field.setAccessible(Field.java:172)
at com.alibaba.com.caucho.hessian.io.JavaDeserializer.getFieldMap(JavaDeserializer.java:340)
at com.alibaba.com.caucho.hessian.io.JavaDeserializer.<init>(JavaDeserializer.java:80)
at com.alibaba.com.caucho.hessian.io.StackTraceElementDeserializer.<init>(StackTraceElementDeserializer.java:56)
at com.alibaba.com.caucho.hessian.io.SerializerFactory.<clinit>(SerializerFactory.java:189)
at org.apache.dubbo.common.serialize.hessian2.Hessian2FactoryManager.createDefaultSerializerFactory(Hessian2FactoryManager.java:81)
at org.apache.dubbo.common.serialize.hessian2.Hessian2FactoryManager.createSerializerFactory(Hessian2FactoryManager.java:77)
at org.apache.dubbo.common.serialize.hessian2.Hessian2FactoryManager.getSerializerFactory(Hessian2FactoryManager.java:62)
at org.apache.dubbo.common.serialize.hessian2.Hessian2ObjectOutput.<init>(Hessian2ObjectOutput.java:44)
at org.apache.dubbo.common.serialize.hessian2.Hessian2Serialization.serialize(Hessian2Serialization.java:57)
at org.apache.dubbo.rpc.protocol.injvm.DefaultParamDeepCopyUtil.copy(DefaultParamDeepCopyUtil.java:45)
at org.apache.dubbo.rpc.protocol.injvm.InjvmInvoker.rebuildValue(InjvmInvoker.java:263)
at org.apache.dubbo.rpc.protocol.injvm.InjvmInvoker.doInvoke(InjvmInvoker.java:172)
at org.apache.dubbo.rpc.protocol.AbstractInvoker.doInvokeAndReturn(AbstractInvoker.java:242)
at org.apache.dubbo.rpc.protocol.AbstractInvoker.invoke(AbstractInvoker.java:186)
at org.apache.dubbo.rpc.listener.ListenerInvokerWrapper.invoke(ListenerInvokerWrapper.java:71)
at org.dromara.common.dubbo.filter.DubboRequestFilter.invoke(DubboRequestFilter.java:42)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.validation.filter.ValidationFilter.invoke(ValidationFilter.java:98)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at com.alibaba.dubbo.rpc.Invoker$CompatibleInvoker.invoke(Invoker.java:75)
at io.seata.integration.dubbo.alibaba.AlibabaDubboTransactionPropagationFilter.invoke(AlibabaDubboTransactionPropagationFilter.java:45)
at com.alibaba.dubbo.rpc.Filter.invoke(Filter.java:31)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.seata.SeataTransactionPropagationConsumerFilter.invoke(SeataTransactionPropagationConsumerFilter.java:53)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.dromara.common.mybatis.filter.DubboDataPermissionFilter.invoke(DubboDataPermissionFilter.java:25)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.metrics.filter.MetricsFilter.invoke(MetricsFilter.java:58)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at cn.dev33.satoken.context.dubbo.filter.SaTokenDubboConsumerFilter.invoke(SaTokenDubboConsumerFilter.java:42)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.invoke(FilterChainBuilder.java:194)
at org.apache.dubbo.rpc.protocol.ReferenceCountInvokerWrapper.invoke(ReferenceCountInvokerWrapper.java:78)
at org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invokeWithContext(AbstractClusterInvoker.java:382)
at org.apache.dubbo.rpc.cluster.support.FailoverClusterInvoker.doInvoke(FailoverClusterInvoker.java:80)
at org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invoke(AbstractClusterInvoker.java:343)
at com.alibaba.csp.sentinel.adapter.dubbo3.SentinelDubboConsumerFilter.syncInvoke(SentinelDubboConsumerFilter.java:82)
at com.alibaba.csp.sentinel.adapter.dubbo3.SentinelDubboConsumerFilter.invoke(SentinelDubboConsumerFilter.java:66)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.router.RouterSnapshotFilter.invoke(RouterSnapshotFilter.java:46)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.monitor.support.MonitorFilter.invoke(MonitorFilter.java:108)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.support.MetricsClusterFilter.invoke(MetricsClusterFilter.java:50)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.protocol.dubbo.filter.FutureFilter.invoke(FutureFilter.java:52)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at com.alibaba.csp.sentinel.adapter.dubbo3.DubboAppContextFilter.invoke(DubboAppContextFilter.java:47)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.support.ObservationSenderFilter.invoke(ObservationSenderFilter.java:61)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.spring.security.filter.ContextHolderParametersSelectedTransferFilter.invoke(ContextHolderParametersSelectedTransferFilter.java:41)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.support.ConsumerClassLoaderFilter.invoke(ConsumerClassLoaderFilter.java:40)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.support.ConsumerContextFilter.invoke(ConsumerContextFilter.java:118)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.invoke(FilterChainBuilder.java:194)
at org.apache.dubbo.rpc.cluster.support.wrapper.AbstractCluster$ClusterFilterInvoker.invoke(AbstractCluster.java:91)
at org.apache.dubbo.rpc.cluster.support.wrapper.ScopeClusterInvoker.invoke(ScopeClusterInvoker.java:149)
at org.apache.dubbo.registry.client.migration.MigrationInvoker.invoke(MigrationInvoker.java:286)
at org.apache.dubbo.rpc.proxy.InvocationUtil.invoke(InvocationUtil.java:57)
at org.apache.dubbo.rpc.proxy.InvokerInvocationHandler.invoke(InvokerInvocationHandler.java:75)
at org.dromara.system.api.RemoteLogServiceDubboProxy3.saveLog(RemoteLogServiceDubboProxy3.java)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77)
at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.base/java.lang.reflect.Method.invoke(Method.java:568)
at org.springframework.aop.support.AopUtils.invokeJoinpointUsingReflection(AopUtils.java:343)
at org.springframework.aop.framework.JdkDynamicAopProxy.invoke(JdkDynamicAopProxy.java:216)
at jdk.proxy2/jdk.proxy2.$Proxy271.saveLog(Unknown Source)
at org.dromara.common.log.event.LogEventListener.saveLog(LogEventListener.java:39)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77)
at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.base/java.lang.reflect.Method.invoke(Method.java:568)
at org.springframework.aop.support.AopUtils.invokeJoinpointUsingReflection(AopUtils.java:343)
at org.springframework.aop.framework.ReflectiveMethodInvocation.invokeJoinpoint(ReflectiveMethodInvocation.java:196)
at org.springframework.aop.framework.ReflectiveMethodInvocation.proceed(ReflectiveMethodInvocation.java:163)
at org.springframework.aop.framework.CglibAopProxy$CglibMethodInvocation.proceed(CglibAopProxy.java:750)
at org.springframework.aop.interceptor.AsyncExecutionInterceptor.lambda$invoke$0(AsyncExecutionInterceptor.java:115)
at java.base/java.util.concurrent.FutureTask.run$$$capture(FutureTask.java:264)
at java.base/java.util.concurrent.FutureTask.run(FutureTask.java)
at java.base/java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:304)
at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)
at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)
at java.base/java.lang.Thread.run(Thread.java:833)
```
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
<!-- What do you expect from the above steps？-->
### Actual Behavior
<!-- What actually happens? -->
If there is an exception, please attach the exception trace:
```
Just put your stack trace here!
```
| https://github.com/apache/dubbo/issues/12545 | https://github.com/apache/dubbo/pull/12553 | 3160602072c198728cca48e449984a4e596e7c37 | a4e9c9efa4ec79ca6e371141762c3da84718c65f | 2023-06-16T01:40:13Z | java | 2023-06-19T01:34:29Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,540 | ["dubbo-cluster/src/main/java/org/apache/dubbo/rpc/cluster/router/condition/config/ProviderAppStateRouter.java"] | dubbo 3.2.2 生产者启动后 会包一大堆警告 This may be caused by condition router get providerApplication is empty | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
100%复现 每次启动都有
* Dubbo version: 3.2.2
* Operating System version: xxx
* Java version: 17
* nacos 2.2.1
* springcloud alibaba 2022-RC2
### Steps to reproduce this issue
1. xxx
2. xxx
3. xxx
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
<!-- What do you expect from the above steps？-->
### Actual Behavior
<!-- What actually happens? -->
If there is an exception, please attach the exception trace:
```
Just put your stack trace here!
```
| https://github.com/apache/dubbo/issues/12540 | https://github.com/apache/dubbo/pull/12729 | 68e624f7643eeb1aa136a382e4432f3b77002f87 | ff60c870edc0878874ae64d1c7c1c5c18a13c519 | 2023-06-15T05:24:26Z | java | 2023-07-15T02:15:38Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,532 | ["dubbo-metadata/dubbo-metadata-api/src/main/java/org/apache/dubbo/metadata/report/identifier/MetadataIdentifier.java", "dubbo-metadata/dubbo-metadata-api/src/main/java/org/apache/dubbo/metadata/report/support/AbstractMetadataReport.java"] | Metrics output unified interface and serviceKey | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
When outputting to the micrometer, uniformly use serviceKey.
| https://github.com/apache/dubbo/issues/12532 | https://github.com/apache/dubbo/pull/12586 | 27d9ce383640359924f98a4ccb66fb6f3f3140d2 | 2c82cf7dfccd435be7f8de7346a937a27f2d0a9a | 2023-06-14T12:18:08Z | java | 2023-06-24T09:02:15Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,526 | ["dubbo-config/dubbo-config-api/src/main/java/org/apache/dubbo/config/ReferenceConfig.java", "dubbo-config/dubbo-config-api/src/main/java/org/apache/dubbo/config/ServiceConfig.java"] | provider注册url上methods参数中会包含重复的方法名 | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.1.11
* Operating System version: MAC
* Java version: 1.8
### Steps to reproduce this issue
1. 使用3.1.11的版本定义一个接口提供服务，接口中包含相同名字不同参数的X个方法，那么在provider的url上methods参数就会有多个重复的方法名
2. 在2.6.x的consumer端引用该服务时，methodInvokerMap会根据相同方法名存放X个重复invoker地址，如果默认使用random负载均衡，相当于该provider有着X倍的权重一样，压力会是X倍
3. 在刚升dubbo3的时候，如果一台机器收到的流量是其他机器的X倍，可能压力较大会有风险
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
在生成provider url的methods参数时，去掉重复的方法名
### Actual Behavior
同名不同参数的方法名，会在url上生成多次
红圈中是灰度发布一台的重复现象
| https://github.com/apache/dubbo/issues/12526 | https://github.com/apache/dubbo/pull/12527 | a613cae2f3d6d42e555575d46c142ef60ec30729 | 467622aac6b62408c1f5861ddec2880cd74e88cc | 2023-06-14T06:05:48Z | java | 2023-06-18T05:45:06Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,493 | ["dubbo-rpc/dubbo-rpc-dubbo/src/main/java/org/apache/dubbo/rpc/protocol/dubbo/DubboInvoker.java", "dubbo-rpc/dubbo-rpc-injvm/src/main/java/org/apache/dubbo/rpc/protocol/injvm/InjvmInvoker.java", "dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/TripleInvoker.java"] | 未及时释放FutureContext造成频繁GC问题 | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.0.12 3.1.8
* Java version: 1.8
### Steps to reproduce this issue
1. consumer及provider 通过环境变量，设置参数 future.sync.set=false
2. provider返回的报文较大
3. 3.0.12版本的调用，dubbo协议不会产生频繁GC问题，triple会产生
4. 3.1.8版本调用，dubbo协议及triple协议都会产生频繁GC问题
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
Consumer通过dubbo或triple请求，设置了future.sync.set=false 不会在IntenalThreadLocal中保存返回结果，不再产生频繁GC问题
<!-- What do you expect from the above steps？-->
### Actual Behavior
统一设置了参数 future.sync.set=false
**3.0.12版本:**
> Dubbo协议调用
及时清除了FutureContext，生效原因
<img width="1135" alt="image" src="https://github.com/apache/dubbo/assets/33999699/cca05d9d-dacd-4cf9-95d9-a72f9e20533d">
并在之后响应过程中清除了FutureContext
<img width="1088" alt="image" src="https://github.com/apache/dubbo/assets/33999699/c63dff14-6814-41ac-8ad2-d35968f4cd8a">
且因参数future.sync.set=false 之后也不再设置FutureContext，从而能够解决频发GC的情况
> Tiple协议调用
Tiple未及时清除了FutureContext原因
<img width="1027" alt="image" src="https://github.com/apache/dubbo/assets/33999699/2a71560a-2053-4d9e-8f88-07e976e7c713">
且没有在接收到响应后清除InternalThreadLocal
虽参数future.sync.set=false，但已经将FutureContext设置进了InternalThreadLocal，会造成频繁GC的情况
**3.1.8版本:**
Dubbo协议、Triple协议都没有及时清除FutureContext
> Dubbo协议调用
<img width="1147" alt="image" src="https://github.com/apache/dubbo/assets/33999699/6b170820-1088-423f-8634-563944f6b00b">
> Tiple协议调用
Triple未清除原因 同3.0.12
<!-- What actually happens? -->
| https://github.com/apache/dubbo/issues/12493 | https://github.com/apache/dubbo/pull/12534 | 8474bf5415c6a7e76f0d9a9682399c9fa34d8a1e | 5d76667808720eec49d6d970a57145142b228634 | 2023-06-11T16:52:51Z | java | 2023-06-30T12:33:17Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,476 | ["dubbo-rpc/dubbo-rpc-api/src/main/java/org/apache/dubbo/rpc/protocol/AbstractProtocol.java", "dubbo-rpc/dubbo-rpc-api/src/main/java/org/apache/dubbo/rpc/protocol/AbstractProxyProtocol.java", "dubbo-rpc/dubbo-rpc-rest/src/main/java/org/apache/dubbo/rpc/protocol/rest/RestProtocol.java"] | java.lang.NoSuchFieldError: serverMap | <img width="739" alt="image" src="https://github.com/apache/dubbo/assets/12210038/ead31890-dc6e-4a82-a4d0-14c323f756cf">
<img width="473" alt="image" src="https://github.com/apache/dubbo/assets/12210038/85708f4f-4989-4324-9fb9-0102b2561927">
<img width="557" alt="image" src="https://github.com/apache/dubbo/assets/12210038/e5965563-cca4-4ade-95e4-d1b4211cdaec">
```
<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 https://maven.apache.org/xsd/maven-4.0.0.xsd">
<modelVersion>4.0.0</modelVersion>
<parent>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-starter-parent</artifactId>
<version>3.1.0</version>
<relativePath/> <!-- lookup parent from repository -->
</parent>
<groupId>com.example</groupId>
<artifactId>provider</artifactId>
<version>0.0.1-SNAPSHOT</version>
<name>provider</name>
<description>provider</description>
<properties>
<java.version>17</java.version>
<dubbo.version>3.2.2</dubbo.version>
</properties>
<dependencies>
<dependency>
<groupId>javax.servlet</groupId>
<artifactId>javax.servlet-api</artifactId>
<version>4.0.1</version>
</dependency>
<dependency>
<groupId>com.caucho</groupId>
<artifactId>hessian</artifactId>
<version>4.0.66</version>
</dependency>
<dependency>
<groupId>org.apache.dubbo.extensions</groupId>
<artifactId>dubbo-rpc-hessian</artifactId>
<version>1.0.1</version>
</dependency>
<dependency>
<groupId>com.dubbo.test</groupId>
<artifactId>api</artifactId>
<version>1.0-SNAPSHOT</version>
</dependency>
<dependency>
<groupId>org.apache.dubbo</groupId>
<artifactId>dubbo-spring-boot-starter</artifactId>
</dependency>
<dependency>
<groupId>org.apache.dubbo</groupId>
<artifactId>dubbo-dependencies-zookeeper-curator5</artifactId>
<type>pom</type>
<exclusions>
<exclusion>
<groupId>org.slf4j</groupId>
<artifactId>slf4j-reload4j</artifactId>
</exclusion>
</exclusions>
</dependency>
<dependency>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-starter</artifactId>
</dependency>
<dependency>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-starter-test</artifactId>
<scope>test</scope>
</dependency>
</dependencies>
<dependencyManagement>
<dependencies>
<dependency>
<groupId>org.apache.dubbo</groupId>
<artifactId>dubbo-spring-boot-starter</artifactId>
<version>${dubbo.version}</version>
</dependency>
<dependency>
<groupId>org.apache.dubbo</groupId>
<artifactId>dubbo-bom</artifactId>
<version>${dubbo.version}</version>
<type>pom</type>
<scope>import</scope>
</dependency>
<dependency>
<groupId>org.apache.dubbo</groupId>
<artifactId>dubbo-dependencies-zookeeper-curator5</artifactId>
<version>${dubbo.version}</version>
<type>pom</type>
</dependency>
</dependencies>
</dependencyManagement>
<build>
<plugins>
<plugin>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-maven-plugin</artifactId>
</plugin>
</plugins>
</build>
</project>
```
| https://github.com/apache/dubbo/issues/12476 | https://github.com/apache/dubbo/pull/12507 | dd94c1561ade672cc253eef79828f6d2daf5a676 | 2cb12ff9a6444b69f6b7f1b3cac7532be93e6083 | 2023-06-07T15:01:54Z | java | 2023-06-12T12:51:13Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,473 | ["dubbo-config/dubbo-config-spring/src/main/java/org/apache/dubbo/config/spring/schema/DubboBeanDefinitionParser.java", "dubbo-config/dubbo-config-spring/src/test/java/org/apache/dubbo/config/spring/schema/DubboNamespaceHandlerTest.java", "dubbo-spring-boot/dubbo-spring-boot-compatible/autoconfigure/src/test/resources/META-INF/spring/dubbo-context.xml"] | Spring bean conflict of dubbo:method | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: xxx
* Operating System version: xxx
* Java version: xxx
### Steps to reproduce this issue
```xml
<bean id="sayHi" class="org.apache.dubbo.samples.provider.GreetingsServiceImpl"/>
<dubbo:service id="GreetingsServiceRef1" interface="org.apache.dubbo.samples.api.GreetingsService" ref="sayHi">
<dubbo:method name="sayHi" timeout="1000" retries="3"/>
</dubbo:service>
```
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
<!-- What do you expect from the above steps？-->
### Actual Behavior
<!-- What actually happens? -->
If there is an exception, please attach the exception trace:
```
Just put your stack trace here!
```
| https://github.com/apache/dubbo/issues/12473 | https://github.com/apache/dubbo/pull/12474 | 04d132fee1cf23f6c0eada7131c9b5ab18b8ded7 | a758a1e7e0efd5d9465cde6088c04f25fa11aa4a | 2023-06-07T08:55:46Z | java | 2023-06-08T04:47:57Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,434 | ["dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/call/BiStreamServerCallListener.java", "dubbo-rpc/dubbo-rpc-triple/src/main/java/org/apache/dubbo/rpc/protocol/tri/transport/TripleServerConnectionHandler.java"] | 使用dubbo triple reactor 方式通信，当调用方程序被杀死停止后，服务提供方建立的流doFinally方法不会收到中断信号？ | 您好，我想咨询下使用dubbo triple reactor 方式通信的程序，当调用方程序被杀死停止后，为什么服务提供方建立的流doFinally方法不会收到中断信号？ 我理想的情况是服务提供方建立的流doFinally方法应该收到中断信号，这样我程序可以做一些资源清理操作，但目前实际程序运行并没有，希望能解答下，谢谢！
环境：
dubbo ：3.1.5
protobuf-java: 3.22.0
rpc.proto文件：
```
syntax = "proto3";
option java_multiple_files = true;
package com.protocol.grpc;
message Request {
int64 id = 1;
string sessionId = 2;
string userId = 3;
string ns = 4;
string cmd = 5;
bool binary = 6;
bytes payload = 7;
}
enum ResponseType {
RESPONSE = 0;
PUSH = 1;
}
message Response {
int64 id = 1;
string sessionId = 2;
string userId = 3;
string ns = 4;
string cmd = 5;
bool binary = 6;
bytes payload = 7;
bool success = 8;
string msg = 9;
ResponseType type = 10;
}
service DataModelService {
rpc connect(stream Request) returns (stream Response);
}
```
下面是简单的提供方代码，通过connect建立连接后，收到消息就原样返回。
```
@Slf4j
@DubboService(group = "test-sse")
@RequiredArgsConstructor
public class SseDataModelServiceProvider extends DubboDataModelServiceTriple.DataModelServiceImplBase {
@Override
public Flux<Response> connect(Flux<Request> request) {
return request.doFinally(signalType -> {
log.info("connect is finally, signalType:{}", signalType);
//TODO
//做一些资源回收
}).doOnNext(requestMsg -> {
log.info("get data: {}", requestMsg);
}).flatMap(requestMsg -> Flux.just(Response
.newBuilder()
.setId(requestMsg.getId())
.setSessionId(requestMsg.getSessionId())
.setUserId(requestMsg.getUserId())
.setBinary(requestMsg.getBinary())
.setType(ResponseType.RESPONSE)
.setCmd(requestMsg.getCmd())
.setNs(requestMsg.getNs())
.setSuccess(true)
.setPayload(ByteString.copyFrom(("payload").getBytes(StandardCharsets.UTF_8))).build())
);
}
}
```
| https://github.com/apache/dubbo/issues/12434 | https://github.com/apache/dubbo/pull/12451 | 00d8a46cd54dff3687f32ade67c0bcddc3285e07 | eba976fec63eb1ccbedd5751162e2f2a888391df | 2023-05-31T07:30:02Z | java | 2023-06-07T03:03:26Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,415 | ["dubbo-common/src/main/java/org/apache/dubbo/common/utils/ToStringUtils.java", "dubbo-rpc/dubbo-rpc-api/src/main/java/org/apache/dubbo/rpc/support/AccessLogData.java"] | trile protocol config accesslog="true" lead to StackOverflowError | * Dubbo version: 3.1.5
* Operating System version: xxx
* Java version: jdk8
go client invoke java,Exception not print on java side,only go client side , AccessLogFilter convert request param (proto generated java class) to json,it lead to cycle invoke:
| https://github.com/apache/dubbo/issues/12415 | https://github.com/apache/dubbo/pull/12427 | f60ac914f463034df4a50e6f871ddfb146c3b541 | 74315babbeb4513c8b4ded4b40b0c254060ae3ca | 2023-05-29T03:34:59Z | java | 2023-06-03T06:47:23Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,359 | ["dubbo-registry/dubbo-registry-api/src/main/java/org/apache/dubbo/registry/integration/RegistryDirectory.java"] | Error exception prompt when the protocol does not exist | ### Environment
* Dubbo version: 3.2.1
* Operating System version: mac m1
* Java version: 1.8
### Steps to reproduce this issue
1. Pull the latest source code of dubbo3.2
2. Start the demo provider project, dubbo-springboot-demo-provider
3. Start the demo comsumer project, dubbo-springboot-demo-comsuer。Configure as follow:
```
dubbo.consumer.protocol: unknown //does not exist
```
### Expected Behavior
Comuser fails to start and reports that the protocol does not exist
### Actual Behavior
Consumer failed to start and printed an unrelated exception
```
[20/05/23 23:43:14:894 CST] main ERROR deploy.DefaultModuleDeployer: [DUBBO] Model start failed: Dubbo Module[1.1.1] start failed: java.lang.IllegalStateException: Failed to check the status of the service org.apache.dubbo.springboot.demo.DemoService. No provider available for the service org.apache.dubbo.springboot.demo.DemoService from the url fgh://192.168.1.105/org.apache.dubbo.springboot.demo.DemoService?application=dubbo-springboot-demo-consumer&background=false&dubbo=2.0.2&executor-management-mode=isolation&file-cache=true&interface=org.apache.dubbo.springboot.demo.DemoService&methods=sayHello,sayHelloAsync&pid=24800&protocol=fgh&qos.enable=true&qos.port=33333®ister.ip=192.168.1.105&release=&side=consumer&sticky=false×tamp=1684597394299&unloadClusterRelated=false to the consumer 192.168.1.105 use dubbo version , dubbo version: , current host: 192.168.1.105, error code: 5-14. This may be caused by , go to https://dubbo.apache.org/faq/5/14 to find instructions.
java.lang.IllegalStateException: Failed to check the status of the service org.apache.dubbo.springboot.demo.DemoService. No provider available for the service org.apache.dubbo.springboot.demo.DemoService from the url fgh://192.168.1.105/org.apache.dubbo.springboot.demo.DemoService?application=dubbo-springboot-demo-consumer&background=false&dubbo=2.0.2&executor-management-mode=isolation&file-cache=true&interface=org.apache.dubbo.springboot.demo.DemoService&methods=sayHello,sayHelloAsync&pid=24800&protocol=fgh&qos.enable=true&qos.port=33333®ister.ip=192.168.1.105&release=&side=consumer&sticky=false×tamp=1684597394299&unloadClusterRelated=false to the consumer 192.168.1.105 use dubbo version
at org.apache.dubbo.config.ReferenceConfig.checkInvokerAvailable(ReferenceConfig.java:647)
at org.apache.dubbo.config.ReferenceConfig.init(ReferenceConfig.java:311)
at org.apache.dubbo.config.ReferenceConfig.get(ReferenceConfig.java:232)
at org.apache.dubbo.config.utils.SimpleReferenceCache.destroyReference(SimpleReferenceCache.java:265)
at org.apache.dubbo.config.utils.SimpleReferenceCache.destroy(SimpleReferenceCache.java:218)
at org.apache.dubbo.config.utils.SimpleReferenceCache.destroy(SimpleReferenceCache.java:242)
at org.apache.dubbo.config.deploy.DefaultModuleDeployer.lambda$referServices$6(DefaultModuleDeployer.java:444)
at java.util.concurrent.ConcurrentHashMap$ValuesView.forEach(ConcurrentHashMap.java:4707)
at org.apache.dubbo.config.deploy.DefaultModuleDeployer.referServices(DefaultModuleDeployer.java:419)
at org.apache.dubbo.config.deploy.DefaultModuleDeployer.startSync(DefaultModuleDeployer.java:173)
at org.apache.dubbo.config.deploy.DefaultModuleDeployer.start(DefaultModuleDeployer.java:145)
at org.apache.dubbo.config.spring.context.DubboDeployApplicationListener.onContextRefreshedEvent(DubboDeployApplicationListener.java:113)
at org.apache.dubbo.config.spring.context.DubboDeployApplicationListener.onApplicationEvent(DubboDeployApplicationListener.java:102)
at org.apache.dubbo.config.spring.context.DubboDeployApplicationListener.onApplicationEvent(DubboDeployApplicationListener.java:47)
at org.springframework.context.event.SimpleApplicationEventMulticaster.doInvokeListener(SimpleApplicationEventMulticaster.java:176)
at org.springframework.context.event.SimpleApplicationEventMulticaster.invokeListener(SimpleApplicationEventMulticaster.java:169)
at org.springframework.context.event.SimpleApplicationEventMulticaster.multicastEvent(SimpleApplicationEventMulticaster.java:143)
at org.springframework.context.support.AbstractApplicationContext.publishEvent(AbstractApplicationContext.java:421)
at org.springframework.context.support.AbstractApplicationContext.publishEvent(AbstractApplicationContext.java:378)
at org.springframework.context.support.AbstractApplicationContext.finishRefresh(AbstractApplicationContext.java:940)
at org.springframework.context.support.AbstractApplicationContext.refresh(AbstractApplicationContext.java:586)
at org.springframework.boot.SpringApplication.refresh(SpringApplication.java:731)
at org.springframework.boot.SpringApplication.refreshContext(SpringApplication.java:408)
at org.springframework.boot.SpringApplication.run(SpringApplication.java:307)
at org.springframework.boot.SpringApplication.run(SpringApplication.java:1303)
at org.springframework.boot.SpringApplication.run(SpringApplication.java:1292)
at org.apache.dubbo.springboot.demo.consumer.ConsumerApplication.main(ConsumerApplication.java:38)
``` | https://github.com/apache/dubbo/issues/12359 | https://github.com/apache/dubbo/pull/12361 | 3fa518505fc54be59e4306300891e54db4610240 | bde608ec3cf0cad32e2704e89e008a83bc27c46c | 2023-05-20T16:06:48Z | java | 2023-06-12T12:47:02Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,353 | ["dubbo-remoting/dubbo-remoting-netty4/src/main/java/org/apache/dubbo/remoting/transport/netty4/NettyChannel.java", "dubbo-remoting/dubbo-remoting-netty4/src/main/java/org/apache/dubbo/remoting/transport/netty4/NettyConfigOperator.java"] | 端口协议复用且encodeInIOThread=false，服务端不能正常响应客户端 | ### Environment
* Dubbo version: 3.2.0
* Operating System version: macos10.15.7
* Java version: 1.8
### Steps to reproduce this issue
服务提供者
```java
public class ApiProvider {
public static void main(String[] args) {
ServiceConfig<GreeterService> serviceConfig = new ServiceConfig<>();
serviceConfig.setInterface(GreeterService.class);
serviceConfig.setRef(new GreeterServiceImpl());
ProtocolConfig protocol = new ProtocolConfig("tri", 20881);
protocol.setExtProtocol("dubbo");
ApplicationConfig applicationConfig = new ApplicationConfig("dubbo-demo-triple-api-provider");
applicationConfig.setRegisterMode("instance");
applicationConfig.setMetadataType("remote");
applicationConfig.setEnableFileCache(false);
DubboBootstrap bootstrap = DubboBootstrap.getInstance();
bootstrap.application(applicationConfig)
.registry(new RegistryConfig("zookeeper://127.0.0.1:2181"))
.protocol(protocol)
.service(serviceConfig)
.start()
.await();
}
}
```
服务消费者
```java
public class ApiConsumer {
public static void main(String[] args) throws InterruptedException {
System.setProperty("dubbo.application.service-discovery.migration", "FORCE_APPLICATION");
ReferenceConfig<GreeterService> referenceConfig = new ReferenceConfig<>();
referenceConfig.setInterface(GreeterService.class);
referenceConfig.setCheck(false);
referenceConfig.setTimeout(3000);
referenceConfig.setProtocol(CommonConstants.DUBBO);
referenceConfig.setRetries(0);
ApplicationConfig applicationConfig = new ApplicationConfig("dubbo-demo-triple-api-consumer");
applicationConfig.setMetadataType("remote");
applicationConfig.setEnableFileCache(false);
DubboBootstrap bootstrap = DubboBootstrap.getInstance();
bootstrap.application(applicationConfig)
.registry(new RegistryConfig("zookeeper://127.0.0.1:2181"))
.protocol(new ProtocolConfig(CommonConstants.DUBBO, -1))
.reference(referenceConfig)
.start();
referenceConfig.get().sayHello(HelloRequest.newBuilder().setName("123").build());
}
}
```
### Expected Behavior
服务端正常响应客户端
### Actual Behavior
服务消费者stacktrace
```
[19/05/23 03:10:58:058 CST] DubboShutdownHook INFO deploy.DefaultModuleDeployer: [DUBBO] Dubbo Module[1.1.1] is stopping., dubbo version: 3.2.0, current host: 10.55.38.81
Exception in thread "main" org.apache.dubbo.rpc.RpcException: Failed to invoke the method sayHello in the service org.apache.dubbo.demo.GreeterService. Tried 1 times of the providers [10.55.38.81:20881] (1/1) from the registry 127.0.0.1:2181 on the consumer 10.55.38.81 using the dubbo version 3.2.0. Last error is: Invoke remote method timeout. method: sayHello, provider: DefaultServiceInstance{serviceName='dubbo-demo-triple-api-provider', host='10.55.38.81', port=20881, enabled=true, healthy=true, metadata={dubbo.endpoints=[{"port":20881,"protocol":"dubbo"},{"port":20881,"protocol":"tri"}], dubbo.metadata.revision=0261e84951c0c038110acb6ff1edbb32, dubbo.metadata.storage-type=remote, timestamp=1684480151231}}, service{name='org.apache.dubbo.demo.GreeterService',group='null',version='null',protocol='dubbo',port='20881',params={executor-management-mode=isolation, side=provider, file-cache=false, release=3.2.0, methods=sayHello,sayHelloAsync,sayMultiHello1,sayMultiHello2, deprecated=false, dubbo=2.0.2, interface=org.apache.dubbo.demo.GreeterService, service-name-mapping=true, register-mode=instance, generic=false, ispuserver=true, metadata-type=remote, application=dubbo-demo-triple-api-provider, prefer.serialization=fastjson2,hessian2, background=false, dynamic=true, anyhost=true},}, cause: org.apache.dubbo.remoting.TimeoutException: Waiting server-side response timeout by scan timer. start time: 2023-05-19 15:10:55.126, end time: 2023-05-19 15:10:58.173, client elapsed: 124 ms, server elapsed: 2923 ms, timeout: 3000 ms, request: Request [id=0, version=2.0.2, twoWay=true, event=false, broken=false, mPayload=0, data=RpcInvocation [methodName=sayHello, parameterTypes=[class org.apache.dubbo.demo.hello.HelloRequest]]], channel: /10.55.38.81:56152 -> /10.55.38.81:20881
at org.apache.dubbo.rpc.cluster.support.FailoverClusterInvoker.doInvoke(FailoverClusterInvoker.java:115)
at org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invoke(AbstractClusterInvoker.java:341)
at org.apache.dubbo.rpc.cluster.router.RouterSnapshotFilter.invoke(RouterSnapshotFilter.java:46)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.monitor.support.MonitorFilter.invoke(MonitorFilter.java:101)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.support.MetricsClusterFilter.invoke(MetricsClusterFilter.java:51)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.protocol.dubbo.filter.FutureFilter.invoke(FutureFilter.java:52)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.support.ObservationSenderFilter.invoke(ObservationSenderFilter.java:61)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.support.ConsumerClassLoaderFilter.invoke(ConsumerClassLoaderFilter.java:40)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.support.ConsumerContextFilter.invoke(ConsumerContextFilter.java:118)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.invoke(FilterChainBuilder.java:194)
at org.apache.dubbo.rpc.cluster.support.wrapper.AbstractCluster$ClusterFilterInvoker.invoke(AbstractCluster.java:91)
at org.apache.dubbo.rpc.cluster.support.wrapper.MockClusterInvoker.invoke(MockClusterInvoker.java:103)
at org.apache.dubbo.rpc.cluster.support.wrapper.ScopeClusterInvoker.invoke(ScopeClusterInvoker.java:131)
at org.apache.dubbo.registry.client.migration.MigrationInvoker.invoke(MigrationInvoker.java:286)
at org.apache.dubbo.rpc.proxy.InvocationUtil.invoke(InvocationUtil.java:57)
at org.apache.dubbo.rpc.proxy.InvokerInvocationHandler.invoke(InvokerInvocationHandler.java:75)
at org.apache.dubbo.demo.GreeterServiceDubboProxy0.sayHello(GreeterServiceDubboProxy0.java)
at org.apache.dubbo.demo.consumer.ApiConsumer.main(ApiConsumer.java:58)
Caused by: java.util.concurrent.ExecutionException: org.apache.dubbo.remoting.TimeoutException: Waiting server-side response timeout by scan timer. start time: 2023-05-19 15:10:55.126, end time: 2023-05-19 15:10:58.173, client elapsed: 124 ms, server elapsed: 2923 ms, timeout: 3000 ms, request: Request [id=0, version=2.0.2, twoWay=true, event=false, broken=false, mPayload=0, data=RpcInvocation [methodName=sayHello, parameterTypes=[class org.apache.dubbo.demo.hello.HelloRequest]]], channel: /10.55.38.81:56152 -> /10.55.38.81:20881
at java.util.concurrent.CompletableFuture.reportGet(CompletableFuture.java:357)
at java.util.concurrent.CompletableFuture.get(CompletableFuture.java:1915)
at org.apache.dubbo.rpc.AsyncRpcResult.get(AsyncRpcResult.java:208)
at org.apache.dubbo.rpc.protocol.AbstractInvoker.waitForResultIfSync(AbstractInvoker.java:286)
at org.apache.dubbo.rpc.protocol.AbstractInvoker.invoke(AbstractInvoker.java:189)
at org.apache.dubbo.rpc.listener.ListenerInvokerWrapper.invoke(ListenerInvokerWrapper.java:71)
at org.apache.dubbo.metrics.filter.MetricsFilter.invoke(MetricsFilter.java:56)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.invoke(FilterChainBuilder.java:194)
at org.apache.dubbo.rpc.protocol.ReferenceCountInvokerWrapper.invoke(ReferenceCountInvokerWrapper.java:78)
at org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invokeWithContext(AbstractClusterInvoker.java:380)
at org.apache.dubbo.rpc.cluster.support.FailoverClusterInvoker.doInvoke(FailoverClusterInvoker.java:81)
... 24 more
Caused by: org.apache.dubbo.remoting.TimeoutException: Waiting server-side response timeout by scan timer. start time: 2023-05-19 15:10:55.126, end time: 2023-05-19 15:10:58.173, client elapsed: 124 ms, server elapsed: 2923 ms, timeout: 3000 ms, request: Request [id=0, version=2.0.2, twoWay=true, event=false, broken=false, mPayload=0, data=RpcInvocation [methodName=sayHello, parameterTypes=[class org.apache.dubbo.demo.hello.HelloRequest]]], channel: /10.55.38.81:56152 -> /10.55.38.81:20881
at org.apache.dubbo.remoting.exchange.support.DefaultFuture.doReceived(DefaultFuture.java:217)
at org.apache.dubbo.remoting.exchange.support.DefaultFuture.received(DefaultFuture.java:181)
at org.apache.dubbo.remoting.exchange.support.DefaultFuture$TimeoutCheckTask.notifyTimeout(DefaultFuture.java:291)
at org.apache.dubbo.remoting.exchange.support.DefaultFuture$TimeoutCheckTask.lambda$run$0(DefaultFuture.java:278)
at org.apache.dubbo.common.threadpool.ThreadlessExecutor$RunnableWrapper.run(ThreadlessExecutor.java:141)
at org.apache.dubbo.common.threadpool.ThreadlessExecutor.waitAndDrain(ThreadlessExecutor.java:70)
at org.apache.dubbo.rpc.AsyncRpcResult.get(AsyncRpcResult.java:202)
... 33 more
```
客户端建立连接，服务端创建NettyChannel时，走主协议tri，创建内部Codec，即org.apache.dubbo.remoting.api.pu.DefaultCodec，是个空实现。
```java
@Override
public void channelActive(ChannelHandlerContext ctx) throws Exception {
super.channelActive(ctx);
NettyChannel channel = NettyChannel.getOrAddChannel(ctx.channel(), url, handler);// 创建nettychannel，内部codec为tri协议获取到DefaultCodec2
}
```
客户端协议确定以后，虽然org.apache.dubbo.remoting.transport.netty4.NettyConfigOperator#configChannelHandler将pipeline中的codec替换为DubboCountCodec了。
```java
if (!(codec2 instanceof DefaultCodec)){
NettyCodecAdapter codec = new NettyCodecAdapter(codec2, channel.getUrl(), handler);
((NettyChannel) channel).getNioChannel().pipeline().addLast(
codec.getDecoder()
).addLast(
codec.getEncoder()
);
}
```
但是org.apache.dubbo.remoting.transport.netty4.NettyChannel#send优化走业务线程encode，NettyChannel内部的codec是空实现，导致最后提交到netty的outputMessage是个空buffer，最终客户端收不到服务端的响应，超时结束。
```java
public void send(Object message, boolean sent) throws RemotingException {
Object outputMessage = message;
if (!encodeInIOThread) {
ByteBuf buf = channel.alloc().buffer();
ChannelBuffer buffer = new NettyBackedChannelBuffer(buf);
codec.encode(this, buffer, message); // 空实现
outputMessage = buf; // 空buffer
}
``` | https://github.com/apache/dubbo/issues/12353 | https://github.com/apache/dubbo/pull/12355 | 08727e8e270ce355a51417d9604dfd2954b6fe6b | 1671332337caffe68d61a3afb47bc4277c7601ee | 2023-05-19T07:23:04Z | java | 2023-05-20T05:54:29Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,334 | ["dubbo-cluster/src/main/java/org/apache/dubbo/rpc/cluster/support/wrapper/ScopeClusterInvoker.java", "dubbo-cluster/src/test/java/org/apache/dubbo/rpc/cluster/support/wrapper/ScopeClusterInvokerTest.java"] | Dubbo 3.2.0 broadcast + injvm not work as expected | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.2.0 / 3.2.1
* Operating System version: Windows 10
* Java version: OpenJDK 17.0.6
### Steps to reproduce this issue
```java
public interface RemoteCacheSyncService {
void flushCache();
}
@DubboService
public class RemoteCacheSyncServiceImpl implements RemoteCacheSyncService {
@Override
public void flushCache() {
// some code
}
}
```
We have 3 microservices, referred to as `A`, `B`, and `C`.
Each service contains the above code.
But, just `C` has a `@DubboReference` to the interface only.
```java
@DubboReference(cluster = ClusterRules.BROADCAST)
RemoteCacheSyncService remoteCacheSyncService;
```
### Expected Behavior
Call **ALL** implementations in **ALL** microservices.
### Actual Behavior
However, since there is a service implementation of the interface in the **same application** ( `C` ),
Dubbo will use the `injvm` protocol. Only the service implementation located in `C` will be invoked.
Same issue with https://github.com/apache/dubbo/issues/6842 , however, it came back again.
| https://github.com/apache/dubbo/issues/12334 | https://github.com/apache/dubbo/pull/12347 | c22db81a8bd26fb1a1cd0e41b04cb0bbe421cec5 | c76e97dc616509fe6f6b2bef3ab922642af7cd06 | 2023-05-16T13:52:18Z | java | 2023-05-19T07:31:54Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,326 | ["dubbo-kubernetes/src/main/java/org/apache/dubbo/registry/kubernetes/util/KubernetesConfigUtils.java"] | k8s client http2Disable default value is wrong | ### Environment
* Dubbo version: 3.2.0
* Operating System version: macos10.15.7
* Java version: 1.8
### Steps to reproduce this issue
### Expected Behavior
set http2Disable default value into http2Disable.
```java
public class KubernetesConfigUtils {
public static Config createKubernetesConfig(URL url) {
Config base = Config.autoConfigure(null);
return new ConfigBuilder(base) //
// ...
.withTrustCerts(url.getParameter(TRUST_CERTS, base.isTrustCerts())) //
.withHttp2Disable(url.getParameter(HTTP2_DISABLE, base.isHttp2Disable())) //
.build();
}
}
```
### Actual Behavior
set trustCerts default value into http2Disable.
```java
public class KubernetesConfigUtils {
public static Config createKubernetesConfig(URL url) {
Config base = Config.autoConfigure(null);
return new ConfigBuilder(base) //
// ...
.withTrustCerts(url.getParameter(TRUST_CERTS, base.isTrustCerts())) //
.withHttp2Disable(url.getParameter(HTTP2_DISABLE, base.isTrustCerts())) // bug
.build();
}
}
```
| https://github.com/apache/dubbo/issues/12326 | https://github.com/apache/dubbo/pull/12328 | a0e43cbcfe5fb8fd54635ef93adb08bb7fa96a51 | 1511ca09912492309bc2a093863eb969223f84b2 | 2023-05-16T09:27:12Z | java | 2023-05-18T08:20:30Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,321 | ["dubbo-config/dubbo-config-api/src/main/java/org/apache/dubbo/config/deploy/DefaultApplicationDeployer.java", "dubbo-config/dubbo-config-api/src/test/java/org/apache/dubbo/config/deploy/DefaultApplicationDeployerTest.java", "dubbo-metrics/dubbo-metrics-default/src/main/java/org/apache/dubbo/metrics/collector/DefaultMetricsCollector.java"] | Dubbo3.2.0启动报错 | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.2.0
* Operating System version: Windows 10
* Java version: 17
### Steps to reproduce this issue
版本 spring boot 3.0.6 启动报错
Pls. provide [GitHub address] to reproduce this issue.
### Expected Behavior
正常启动
### Actual Behavior
<!-- What actually happens? -->
If there is an exception, please attach the exception trace:
java.lang.NoClassDefFoundError: io/prometheus/client/exporter/HttpConnectionFactory
at org.apache.dubbo.metrics.prometheus.PrometheusMetricsReporterFactory.createMetricsReporter(PrometheusMetricsReporterFactory.java:43)
at org.apache.dubbo.metrics.report.MetricsReporterFactory$Adaptive.createMetricsReporter(MetricsReporterFactory$Adaptive.java)
at org.apache.dubbo.config.deploy.DefaultApplicationDeployer.initMetricsReporter(DefaultApplicationDeployer.java:392)
at org.apache.dubbo.config.deploy.DefaultApplicationDeployer.initialize(DefaultApplicationDeployer.java:216)
at org.apache.dubbo.config.deploy.DefaultModuleDeployer.prepare(DefaultModuleDeployer.java:488)
at org.apache.dubbo.config.spring.context.DubboConfigApplicationListener.initDubboConfigBeans(DubboConfigApplicationListener.java:73)
at org.apache.dubbo.config.spring.context.DubboConfigApplicationListener.onApplicationEvent(DubboConfigApplicationListener.java:59)
at org.apache.dubbo.config.spring.context.DubboConfigApplicationListener.onApplicationEvent(DubboConfigApplicationListener.java:37)
at org.springframework.context.event.SimpleApplicationEventMulticaster.doInvokeListener(SimpleApplicationEventMulticaster.java:176)
at org.springframework.context.event.SimpleApplicationEventMulticaster.invokeListener(SimpleApplicationEventMulticaster.java:169)
at org.springframework.context.event.SimpleApplicationEventMulticaster.multicastEvent(SimpleApplicationEventMulticaster.java:143)
at org.springframework.context.event.SimpleApplicationEventMulticaster.multicastEvent(SimpleApplicationEventMulticaster.java:131)
at org.springframework.context.support.AbstractApplicationContext.registerListeners(AbstractApplicationContext.java:880)
at org.springframework.context.support.AbstractApplicationContext.refresh(AbstractApplicationContext.java:581)
at org.springframework.boot.web.servlet.context.ServletWebServerApplicationContext.refresh(ServletWebServerApplicationContext.java:146)
at org.springframework.boot.SpringApplication.refresh(SpringApplication.java:732)
at org.springframework.boot.SpringApplication.refreshContext(SpringApplication.java:434)
at org.springframework.boot.SpringApplication.run(SpringApplication.java:310)
at org.springframework.boot.SpringApplication.run(SpringApplication.java:1304)
at org.springframework.boot.SpringApplication.run(SpringApplication.java:1293)
at com.crown.amish.user.UserApplication.main(UserApplication.java:17)
Caused by: java.lang.ClassNotFoundException: io.prometheus.client.exporter.HttpConnectionFactory
at java.base/jdk.internal.loader.BuiltinClassLoader.loadClass(BuiltinClassLoader.java:641)
at java.base/jdk.internal.loader.ClassLoaders$AppClassLoader.loadClass(ClassLoaders.java:188)
at java.base/java.lang.ClassLoader.loadClass(ClassLoader.java:520)
... 21 common frames omitted
Just put your stack trace here!
```
| https://github.com/apache/dubbo/issues/12321 | https://github.com/apache/dubbo/pull/12349 | 6bd2174f01ef266c652845da5d7e520b6e486df6 | c22db81a8bd26fb1a1cd0e41b04cb0bbe421cec5 | 2023-05-16T03:44:12Z | java | 2023-05-19T07:30:18Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,320 | ["README.md"] | Some Flaky tests | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [X] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: xxx
* Operating System version: ubuntu 22.04
* Java version: 11
### Comments
Hello, We tried running your project and discovered that it contains some flaky tests (i.e., tests that nondeterministically pass and fail). We found these tests to fail more frequently when running them on certain machines of ours.
To prevent others from running this project and its tests in machines that may result in flaky tests, we suggest adding information to the README.md file indicating the minimum resource configuration for running the tests of this project as to prevent observation of test flakiness.
If we run this project in a machine with 1cpu and 1gb ram, we observe flaky tests. We found that the tests in this project did not have any flaky tests when we ran it on machines with 2cpu and 2gb ram.
Here is a list of the tests we have identified and their likelihood of failure on a system with less than the recommended 2 CPUs and 2 GB RAM.
1. org.apache.dubbo.rpc.cluster.support.MergeableClusterInvokerTest#testGetMenuSuccessfully (4 out 10)
2. org.apache.dubbo.rpc.protocol.dubbo.RpcFilterTest#testRpcFilter (1 out 10)
3. org.apache.dubbo.rpc.filter.ExecuteLimitFilterTest#testMoreThanExecuteLimitInvoke ( 2 out 10)
4. org.apache.dubbo.rpc.filter.ActiveLimitFilterTest#testInvokeGreaterActives ( 1 out 10)
5. org.apache.dubbo.rpc.protocol.rest.SpringMvcRestProtocolTest#testRestProtocol ( 7 out 10)
6. org.apache.dubbo.rpc.protocol.rest.SpringMvcRestProtocolTest#testRestProtocolWithContextPath ( 7 out 10)
7. org.apache.dubbo.rpc.protocol.dubbo.MultiThreadTest#testDubboMultiThreadInvoke ( 4 out 10)
Please let me know if you would like us to create a pull request on this matter (possibly to the readme of this project).
Thank you for your attention to this matter. We hope that our recommendations will be helpful in improving the quality and performance of your project, especially for others to use.
## Reproducing
~~~Dockerfile
FROM maven:3.5.4-jdk-11
WORKDIR /home/
RUN git clone https://github.com/apache/dubbo && \
cd dubbo && \
git checkout acd4212f597813411b6634377f96ac2bb3888a07
WORKDIR /home/dubbo
RUN mvn install -DskipTests
ENTRYPOINT ["mvn", "test", "-fn"]
~~~
Build the image:
~~~bash
$> mkdir tmp
$> cp Dockerfile tmp
$> cd tmp
$> docker build -t dubbo . # estimated time of build 3m
~~~
Running:
this configuration likely prevents flakiness (no flakiness in 10 runs)
~~~bash
$> docker run --rm --memory=2g --cpus=2 --memory-swap=-1 dubbo | tee output.txt
$> grep "Failures:" output.txt # checking results
~~~
this other configuration –similar to the previous– can’t prevent flaky tests (observation in 10 runs)
~~~bash
$> docker run --rm --memory=1g --cpus=1 --memory-swap=-1 dubbo | tee output2.txt
$> grep "Failures:" output2.txt # checking results
~~~
| https://github.com/apache/dubbo/issues/12320 | https://github.com/apache/dubbo/pull/12391 | 9eec5973fc5261c8ad08103aee9e16e5313c3e16 | c7bc74557cfa037d6209b10764233cf1ce32eef8 | 2023-05-15T18:59:10Z | java | 2023-05-29T06:48:24Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,315 | ["dubbo-config/dubbo-config-api/src/main/java/org/apache/dubbo/config/deploy/DefaultModuleDeployer.java"] | DefaultModuleDeployer start fail, but export MetadataService | ### Environment
* Dubbo version: 3.2.0
* Operating System version: macos10.15.7
* Java version: 1.8
### Steps to reproduce this issue
使用各种方式导致Module启动失败。
### Expected Behavior
Module启动失败，不暴露MetadataService。
即不执行DefaultApplicationDeployer#prepareApplicationInstance。
### Actual Behavior
由于org.apache.dubbo.config.deploy.DefaultModuleDeployer#onModuleFailed传入错误的状态STARTED，导致执行DefaultApplicationDeployer#prepareApplicationInstance。
```java
private void onModuleFailed(String msg, Throwable ex) {
try {
setFailed(ex);
logger.error(CONFIG_FAILED_START_MODEL, "", "", "Model start failed: " + msg, ex);
applicationDeployer.notifyModuleChanged(moduleModel, DeployState.STARTED);
} finally {
completeStartFuture(false);
}
}
```
| https://github.com/apache/dubbo/issues/12315 | https://github.com/apache/dubbo/pull/12316 | 1511ca09912492309bc2a093863eb969223f84b2 | 5fb15a130d8f3b5503c93afbeb9892eaa0238f7a | 2023-05-15T08:51:26Z | java | 2023-05-18T09:45:14Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,290 | ["dubbo-plugin/dubbo-qos/src/main/java/org/apache/dubbo/qos/command/impl/InvokeTelnet.java"] | the telnet invoke did not clean the ThreadLocal, causing the request response pairing errors | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [x] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.2.0, 2.7.x, etc
* Operating System version: Windows 10（irrelevant）
* Java version: java 8（irrelevant）
### Steps to reproduce this issue
1. write the interface implementation class on the provider side, use RpcContext.startAsync() for asynchronous operations.
2. On the consumer side, call this interface in a loop and verify whether the returned result corresponds to the requested parameter.
3. Then, use the telnet invoke command to call this interface and observe whether the consumer side encounters an incorrect request and result pairing.
4. Then close the consumer, and observe whether the AsyncContext value returned by RpcContext.startAsync() on the provider side is the same object multiple times during the telnet invoke execution.
### Expected Behavior
1. the consumer always get the right response
2. RpcContext.startAsync() returns different instance per each request
### Actual Behavior
1. the consumer get mismatch response（only appeared in versions before 3.x）
2. RpcContext.startAsync() returns the same object in multiple requests
---
When using RpcContext.startAsync() to return results asynchronously on the provider side, the context (RpcContext, AsyncContext, etc.) is stored in ThreadLocal by default. Under normal circumstances, the ContextFilter will clear the data in ThreadLocal. However, when the request is triggered by telnet invoke, ContextFilter will not be executed, resulting in ThreadLocal not being cleaned up. When this thread handles the next request, it obtains the context of the previous request, leading to problems.
In versions before 3.x, this can cause consumers to get the wrong data. For example, request B gets the result of request A (same interface, same method), or a request for calling interface C gets the result of interface D (different interface, different method).
In the 3.x version, since the invoke command is assigned to QoS and executed in a separate thread pool, therefore, there will be no issue of response pairing errors with normal consumer requests(if they can be configured to share the same thread pool later, there is still this risk). However, it is still unreasonable to not clean up ThreadLocal after use.
In addition, telnet invoke cannot obtain the results asynchronously, which brings trouble to the operation and maintenance. Therefore, it can be optimized to obtain the results synchronously, making it easier to troubleshoot problems through telnet invoke.
当在provider端使用RpcContext.startAsync()进行异步返回结果时，会使用ThreadLocal存放上下文（RpcContext、AsyncContext等），正常情况下，会由ContextFilter清理ThreadLocal中的数据，当请求是由telnet invoke触发时，不会执行到ContextFilter，导致ThreadLocal未被清理，当此线程去处理下一个请求时，拿到的上下文是上一次请求的结果，从而引发问题。
在3.x以前的版本中，会造成正常的consumer拿到错误的数据，例如B请求拿到的是A请求的结果（同接口同方法），甚至是调用C接口的请求拿到D接口的结果（不同接口不同方法）。
在3.x版本中，由于invoke命令被划分到qos，并且是在单独的线程池中执行，因此不会与正常的consumer的请求发生串数据的情况（如果后续可以通过配置让两者共用线程池，则仍有此风险）。即便如此，ThreadLocal在使用完毕后却不清理仍旧是不合理的。
复现步骤：在provider端编写接口实现类，使用RpcContext.startAsync()进行异步操作；consumer端循环调用此接口，并校验返回结果是否与请求参数对应；然后通过telnet invoke命令调用此接口，观察consumer端是否出现请求与结果不对应的情况，然后关掉consumer，在telnet invoke执行时，观察provider端RpcContext.startAsync()返回的AsyncContext，是否多次为相同的对象。
另外telnet invoke无法拿到异步的结果，给运维时带来麻烦，此处可以优化为同步获取结果，方便通过telnet invoke排查问题
| https://github.com/apache/dubbo/issues/12290 | https://github.com/apache/dubbo/pull/12291 | 7c66d1a968d0388dc3fdbc8e2ff7ee54ffdc2c17 | c916b82ab5fbd2ad9afedbed453c36a2a9865564 | 2023-05-10T16:27:20Z | java | 2023-05-11T11:07:45Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,248 | ["dubbo-rpc/dubbo-rpc-api/src/main/java/org/apache/dubbo/rpc/filter/GenericFilter.java", "dubbo-rpc/dubbo-rpc-api/src/main/java/org/apache/dubbo/rpc/filter/GenericImplFilter.java"] | The generic call in bean mode passes a null parameter and reports an error NPE | ### Environment
* Dubbo version: 3.1
* Operating System version: mac m1
* Java version: 1.8
### Steps to reproduce this issue
1. Use the following code to make a generalization call of the bean mode
2. ```java
public interface DemoService {
String sayHello(String name);
}
```
3. ```java
ReferenceConfig<GenericService> referenceConfig = new ReferenceConfig<>();
referenceConfig.setInterface(DemoService.class.getCanonicalName());
applicationConfig.setRegistry(registryConfig);
referenceConfig.setApplication(applicationConfig);
referenceConfig.setGeneric(GENERIC_SERIALIZATION_BEAN);
GenericService genericService = referenceConfig.get();
JavaBeanDescriptor result = (JavaBeanDescriptor)genericService.$invoke("sayHello", new String[]{"java.lang.String"}, new Object[]{JavaBeanSerializeUtil.serialize(null)});
```
### Expected Behavior
call success
### Actual Behavior
If there is an exception, please attach the exception trace:
```
Caused by: java.lang.NullPointerException
at org.apache.dubbo.rpc.filter.GenericImplFilter.invoke(GenericImplFilter.java:126)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:327)
at org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.invoke(FilterChainBuilder.java:194)
at org.apache.dubbo.rpc.protocol.ReferenceCountInvokerWrapper.invoke(ReferenceCountInvokerWrapper.java:78)
at org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invokeWithContext(AbstractClusterInvoker.java:379)
at org.apache.dubbo.rpc.cluster.support.FailoverClusterInvoker.doInvoke(FailoverClusterInvoker.java:81)
``` | https://github.com/apache/dubbo/issues/12248 | https://github.com/apache/dubbo/pull/12249 | da9fbbdbf8b46a35af9a202f733b5c4d11f6b807 | 838b7c9c4b24b5ab3f93607cca4cd49ef7aed4de | 2023-05-07T07:56:36Z | java | 2023-05-07T23:41:04Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,202 | ["dubbo-remoting/dubbo-remoting-netty4/src/main/java/org/apache/dubbo/remoting/transport/netty4/NettyChannelHandler.java", "dubbo-remoting/dubbo-remoting-netty4/src/main/java/org/apache/dubbo/remoting/transport/netty4/NettyPortUnificationServer.java", "dubbo-remoting/dubbo-remoting-netty4/src/main/java/org/apache/dubbo/remoting/transport/netty4/NettyPortUnificationServerHandler.java"] | dubbo2.7.x升级到3.2.x后 应用跑两三天后就发生内存泄漏 | dubbo2.7.x升级到3.2.x后 应用跑两三天后就发生内存泄漏。
在 org.apache.dubbo.remoting.transport.netty4.NettyChannel的劝酒的私有的静态成员变量CHANNEL_MAP里的。
里面55多万个元素，1.1G的大小。
到后面AbstractInvoker.waitForResultIfSync(底层是threadlessExecutor.waitAndDrain)会卡住，比如： 某个服务正常调用不到100ms就返回了，应用跑两三天后就开始 那个正常的服务调用开始20%的概率出现一次两三秒的卡顿。
| https://github.com/apache/dubbo/issues/12202 | https://github.com/apache/dubbo/pull/12212 | 3552347d44d700b7413650dc279c90a91cf25504 | 830c460c0a2db7ad6d5caed9a0b7e2890952da77 | 2023-04-27T02:19:37Z | java | 2023-04-28T09:14:01Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,153 | [".licenserc.yaml", "NOTICE", "dubbo-metrics/dubbo-metrics-api/src/main/java/org/apache/dubbo/metrics/aggregate/DubboAbstractTDigest.java", "dubbo-metrics/dubbo-metrics-api/src/main/java/org/apache/dubbo/metrics/aggregate/DubboMergingDigest.java", "dubbo-metrics/dubbo-metrics-api/src/main/java/org/apache/dubbo/metrics/aggregate/TimeWindowQuantile.java", "dubbo-metrics/dubbo-metrics-api/src/test/java/org/apache/dubbo/metrics/aggregate/TimeWindowQuantileTest.java", "pom.xml"] | dubbo 3.2 turn on metrics report error ArrayIndexOutOfBoundsException | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
### Environment
* Dubbo version: 3.2.0
* Operating System version: linux
* Java version: jdk 1.8
### Steps to reproduce this issue
1. turn on metrics
metrics:
protocol: prometheus
export-service-port: x0003
enable-jvm: true
enable-threadpool: true
enable-registry: true
enable-metadata: true
2. pressure test 100 thread
If there is an exception, please attach the exception trace:
Caused by: java.lang.ArrayIndexOutOfBoundsException: null
com.tdunning.math.stats.MergingDigest.merge(MergingDigest.java:381)
com.tdunning.math.stats.MergingDigest.mergeNewValues(MergingDigest.java:363)
com.tdunning.math.stats.MergingDigest.mergeNewValues(MergingDigest.java:353)
com.tdunning.math.stats.MergingDigest.add(MergingDigest.java:259)
com.tdunning.math.stats.MergingDigest.add(MergingDigest.java:251)
com.tdunning.math.stats.AbstractTDigest.add(AbstractTDigest.java:129)
org.apache.dubbo.metrics.aggregate.TimeWindowQuantile.add(TimeWindowQuantile.java:51)
org.apache.dubbo.metrics.collector.AggregateMetricsCollector.onRTEvent(AggregateMetricsCollector.java:92)
org.apache.dubbo.metrics.collector.AggregateMetricsCollector.onEvent(AggregateMetricsCollector.java:82)
org.apache.dubbo.metrics.event.SimpleMetricsEventMulticaster.publishEvent(SimpleMetricsEventMulticaster.java:48)
org.apache.dubbo.metrics.collector.sample.MethodMetricsSampler.lambda$rtConfigure$3(MethodMetricsSampler.java:61)
org.apache.dubbo.metrics.collector.sample.SimpleMetricsCountSampler.addRT(SimpleMetricsCountSampler.java:120)
org.apache.dubbo.metrics.filter.MethodMetricsInterceptor.rtTime(MethodMetricsInterceptor.java:97)
org.apache.dubbo.metrics.filter.MethodMetricsInterceptor.onCompleted(MethodMetricsInterceptor.java:101)
org.apache.dubbo.metrics.filter.MethodMetricsInterceptor.afterMethod(MethodMetricsInterceptor.java:58)
org.apache.dubbo.metrics.filter.MetricsFilter.onResponse(MetricsFilter.java:64)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.lambda$invoke$1(FilterChainBuilder.java:218)
org.apache.dubbo.rpc.AsyncRpcResult.lambda$whenCompleteWithContext$0(AsyncRpcResult.java:228)
java.util.concurrent.CompletableFuture.uniWhenComplete(CompletableFuture.java:774)
java.util.concurrent.CompletableFuture.uniWhenCompleteStage(CompletableFuture.java:792)
java.util.concurrent.CompletableFuture.whenComplete(CompletableFuture.java:2153)
org.apache.dubbo.rpc.AsyncRpcResult.whenCompleteWithContext(AsyncRpcResult.java:224)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.invoke(FilterChainBuilder.java:195)
org.apache.dubbo.rpc.protocol.ReferenceCountInvokerWrapper.invoke(ReferenceCountInvokerWrapper.java:78)
org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invokeWithContext(AbstractClusterInvoker.java:380)
org.apache.dubbo.rpc.cluster.support.FailoverClusterInvoker.doInvoke(FailoverClusterInvoker.java:81)
org.apache.dubbo.rpc.cluster.support.AbstractClusterInvoker.invoke(AbstractClusterInvoker.java:341)
org.apache.dubbo.rpc.cluster.router.RouterSnapshotFilter.invoke(RouterSnapshotFilter.java:46)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
org.apache.dubbo.monitor.support.MonitorFilter.invoke$original$vkd7zZMh(MonitorFilter.java:101)
org.apache.dubbo.monitor.support.MonitorFilter.invoke$original$vkd7zZMh$accessor$y8nuUX1o(MonitorFilter.java)
org.apache.dubbo.monitor.support.MonitorFilter$auxiliary$YZgq7bd2.call(Unknown Source)
org.apache.skywalking.apm.agent.core.plugin.interceptor.enhance.InstMethodsInter.intercept(InstMethodsInter.java:86)
org.apache.dubbo.monitor.support.MonitorFilter.invoke(MonitorFilter.java)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
org.apache.dubbo.rpc.cluster.filter.support.MetricsClusterFilter.invoke(MetricsClusterFilter.java:51)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
org.apache.dubbo.rpc.protocol.dubbo.filter.FutureFilter.invoke(FutureFilter.java:52)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
org.apache.dubbo.rpc.cluster.specifyaddress.AddressSpecifyClusterFilter.invoke(AddressSpecifyClusterFilter.java:40)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
org.apache.dubbo.spring.security.filter.ContextHolderParametersSelectedTransferFilter.invoke(ContextHolderParametersSelectedTransferFilter.java:41)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
org.apache.dubbo.rpc.cluster.filter.support.ConsumerClassLoaderFilter.invoke(ConsumerClassLoaderFilter.java:40)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
com.mea.pay.common.dubbo.DubboConsumerContextFilter.invoke(DubboConsumerContextFilter.java:90)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
org.apache.dubbo.rpc.cluster.filter.support.ConsumerContextFilter.invoke(ConsumerContextFilter.java:118)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CopyOfFilterChainNode.invoke(FilterChainBuilder.java:331)
org.apache.dubbo.rpc.cluster.filter.FilterChainBuilder$CallbackRegistrationInvoker.invoke(FilterChainBuilder.java:194)
org.apache.dubbo.rpc.cluster.support.wrapper.AbstractCluster$ClusterFilterInvoker.invoke(AbstractCluster.java:91)
org.apache.dubbo.rpc.cluster.support.wrapper.MockClusterInvoker.invoke(MockClusterInvoker.java:103)
com.mea.pay.common.dubbo.TagInvoker.invoke(TagInvoker.java:45)
org.apache.dubbo.rpc.cluster.support.wrapper.ScopeClusterInvoker.invoke(ScopeClusterInvoker.java:131)
org.apache.dubbo.registry.client.migration.MigrationInvoker.invoke(MigrationInvoker.java:284)
\t... 91 common frames omitted
| https://github.com/apache/dubbo/issues/12153 | https://github.com/apache/dubbo/pull/12223 | c5f8ce64787cbb17628ac5176947d46a457ee0ee | 8b65828767ece40aa54fee1bd155b3f7a6185fac | 2023-04-21T01:41:50Z | java | 2023-05-06T15:02:01Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,132 | ["dubbo-metrics/dubbo-metrics-api/src/main/java/org/apache/dubbo/metrics/model/key/MetricsKey.java", "dubbo-metrics/dubbo-metrics-default/src/main/java/org/apache/dubbo/metrics/collector/AggregateMetricsCollector.java"] | Add P50 and P90 RT metrics | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
## Describe the proposal
<!-- Please use this for a concrete design proposal for functionality. -->
<!-- If you just want to request a new feature and discuss the possible business value, create a Feature Request. -->
| https://github.com/apache/dubbo/issues/12132 | https://github.com/apache/dubbo/pull/12156 | b7fc2b9a6a6e281a3e2fdd32a3a41e83910173e1 | a0e43cbcfe5fb8fd54635ef93adb08bb7fa96a51 | 2023-04-19T06:17:59Z | java | 2023-05-18T07:38:52Z |
closed | apache/dubbo | https://github.com/apache/dubbo | 12,131 | ["dubbo-common/src/main/java/org/apache/dubbo/config/AbstractInterfaceConfig.java", "dubbo-config/dubbo-config-api/src/main/java/org/apache/dubbo/config/deploy/DefaultMetricsServiceExporter.java", "dubbo-metrics/dubbo-metrics-default/src/main/java/org/apache/dubbo/metrics/collector/HistogramMetricsCollector.java"] | Enable aggregation and histogram by default | <!-- If you need to report a security issue please visit https://github.com/apache/dubbo/security/policy -->
- [ ] I have searched the [issues](https://github.com/apache/dubbo/issues) of this repository and believe that this is not a duplicate.
## Describe the proposal
<!-- Please use this for a concrete design proposal for functionality. -->
<!-- If you just want to request a new feature and discuss the possible business value, create a Feature Request. -->
@wxbty PTAL | https://github.com/apache/dubbo/issues/12131 | https://github.com/apache/dubbo/pull/12137 | 3fd0a5a922f42c8914a70d0a89134e9f55782dd9 | 55a8940ad1a98d8ad78da67d8b73856768e09a53 | 2023-04-19T06:17:27Z | java | 2023-06-01T09:19:25Z |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.