You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "XComp (via GitHub)" <gi...@apache.org> on 2023/05/04 07:52:25 UTC
[GitHub] [flink] XComp commented on pull request #22422: [FLINK-31838][runtime] Moves thread handling for leader event listener calls from DefaultMultipleComponentLeaderElectionService into DefaultLeaderElectionService
XComp commented on PR #22422:
URL: https://github.com/apache/flink/pull/22422#issuecomment-1534245422
There's a test failure that popped up in [CI](https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=48461&view=logs&j=af184cdd-c6d8-5084-0b69-7e9c67b35f7a&t=0f3adb59-eefa-51c6-2858-3654d9e0749d&l=15266): Due to the fact that the drivers could trigger the event handling methods (i.e. `onRevokeLeadership` and `onGrantLeadership`) in a separate thread, it could happen that we stop the `DefaultLeaderElectionService` but there's still a event handling call in some thread queue that's triggered after the service is stopped (and consequently, the service's `leadershipOperationExecutor`). That will lead to a `RejectedExecutionException` because we're not checking whether the service is still running before putting the task in the executor's pool.
We actually have unit tests for covering this issue. That just didn't fire because the `DirectExecutorService` implementation doesn't follow the `ExecutorService`'s interface contract. I created FLINK-31995 to cover this issue.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: issues-unsubscribe@flink.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org