Skip to content
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

Revert "Revert "[Java] fix test hang occasionally when running FailureTest (#13934)" (#13992)" #14008

Merged
merged 2 commits into from
Feb 9, 2021

Conversation

kfstorm
Copy link
Member

@kfstorm kfstorm commented Feb 9, 2021

Why are these changes needed?

Revert #13992. For some reason, workers hang when exiting the process. So I skipped std::atexit if it's on Windows. Here is the stack trace:

image

I also saw that test_atexit_handler is skipped on Windows. Maybe they are related.

Related issue number

Checks

  • I've run scripts/format.sh to lint the changes in this PR.
  • I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
  • Testing Strategy
    • Unit tests
    • Release tests
    • This PR is not tested :(

@kfstorm kfstorm requested a review from simon-mo February 9, 2021 13:57
@kfstorm
Copy link
Member Author

kfstorm commented Feb 9, 2021

@edoakes Do you have any insights on this?

@edoakes
Copy link
Contributor

edoakes commented Feb 9, 2021

@kfstorm sorry, not too familiar with std::atexit behavior. Looks like it's an issue with the completion queue cleanup in either the gRPC client or gRPC server though

@simon-mo simon-mo merged commit e0b8179 into ray-project:master Feb 9, 2021
@kfstorm kfstorm deleted the revert_13992 branch February 10, 2021 05:24
fishbone added a commit to fishbone/ray that referenced this pull request Feb 16, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants