E2E test fixes#35414
Conversation
… removing Ignitor)
javiercn
left a comment
There was a problem hiding this comment.
Looks great, thanks for taking on this.
…ulTerminationTests.cs Co-authored-by: Tanay Parikh <TanayParikh@users.noreply.github.com>
|
Thanks for approving, @javiercn and @TanayParikh. I was a bit premature in requesting review on this PR, since I needed to make a bunch of other small changes before it was ready. If either of you wants to review the changes since you saw it, please go ahead, but no need to if you don't feel like it because the changes are pretty minor, test-only, and I think noncontroversial. |
|
Also just for the record, I've run the E2E tests 5 times in a row locally this afternoon, and all 5 times got a 100% pass rate. I'm going to keep running these tests on a loop to collect more data, but at this point it looks like remaining reliability issues are probably CI-specific and not inherent to these tests. |
|
/backport to release/6.0-rc1 |
|
Started backporting to release/6.0-rc1: https://github.com/dotnet/aspnetcore/actions/runs/1143942181 |

As of this PR:
This doesn't address the reliability/consistency concerns, which we still need to take on as a later piece of work. Running the tests locally I don't see many (or any?) random failures of the form "expected x, got y" but I do see random failures like "the test server did not start". This might be assuaged by running without parallelism but it's a whole separate investigation.