-
Notifications
You must be signed in to change notification settings - Fork 4.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
System.Net tests failing with System.TimeoutException #54778
Comments
Tagging subscribers to this area: @dotnet/ncl Issue DetailsFor example:
|
Failed in #54656 Examples:
|
Failed in #54774 Examples:
|
Failures 4/28-6/27 (incl. PRs):
|
Looks like a regression around 6/18-6/23 ... @dotnet/ncl any idea what might have caused the regression? |
@geoffkizer 2 of the tests are HTTP/3 Mock cookies tests -- is it perhaps caused by your PR #54727? |
Yes, these tests were disabled until that PR. Looks like these are still flaky, and apparently much flakier on Linux than Windows for whatever reason. We should disable these again. I'll put up a PR. |
Thanks @geoffkizer! We still have 3 tests not covered by your PR ... we need to figure out what to do about them and likely disable them as well. |
I'm seeing a ton of TimeoutExceptions on my PR to disable these tests. For example: https://helixre8s23ayyeko0k025g8.blob.core.windows.net/dotnet-runtime-refs-pull-54843-merge-5fea1e54014e4d049e/System.Net.Http.Functional.Tests/console.1c5cee98.log?sv=2019-07-07&se=2021-07-18T17%3A51%3A18Z&sr=c&sp=rl&sig=eMv7cbmp3wfyHjUEuXsbeRsXa%2FtVrWqCy%2FPhbk8pVPQ%3D Not sure what is going on here, but it doesn't look isolated to any particular test(s). |
Top 27 methods by number of failures caused by timeout are all happening on
|
Here's breakdown of main branch per date, demonstrating that something happened on 6/23: Failures 3/12-6/29 (incl. PRs) - main branch:
Total 6/23-6/29: 1970 ... out of that 362 are Http3_Mock tests Top hitter tests breakdown:
|
Some of these tests are pretty basic -- single request, no delays or timing issue, etc. Strange. |
Yes, it almost feels like something changed in infra ... |
@geoffkizer I updated the table -- there are 4 tests in *_Http3_Mock category. Could they have regressed due to your recent changes? |
@karelz it's quite easy for me to re-send a job from > 4 days ago to the current machines and see if it still has the problem, even repeatedly if that's useful. Just pick a "blessed" one and I can send it around. Some more thoughts:
|
The mass failures across all PRs started happening somewhere around 9pm on the 24th (+- a few h) - that is why I am suspicious of c88da29). @MattGal Would you be able to run CI on the runtime at 1d84e01 (parent of the above commit)? I've been running CI on a revert of the commit in #54939 and haven't hit a timeout failure yet. |
GetAsync_CancelDuringResponseHeadersReceived_TaskCanceledQuickly is a test I enabled for HTTP3 in my PR. so it's possible that this is a separate regression from whatever else is going on here. But it seems like we should figure out the general issue here first. |
Looks like the failures stopped with #55006. |
There are still some occurrences of test
|
For example:
The text was updated successfully, but these errors were encountered: