[
https://issues.apache.org/jira/browse/FLINK-40446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18115126#comment-18115126
]
Martijn Visser commented on FLINK-40446:
----------------------------------------
[~fanrui] The fix for this is on master as 9c1015353f8 since 2026-08-20, but
the ticket is still open with no fixVersion, and it was not backported.
release-1.20 hits the same race in UnalignedCheckpointRescaleSameUpstreamITCase
in all four Azure nightlies of 2026-09-11 to 2026-09-14, on both
parameterisations:
{code}
java.io.IOException: Failed to delete temp directory
/tmp/junit12972024642090864369. The following paths could not be deleted (see
suppressed exceptions for details):
<jobid>/taskowned/job_<jobid>_tm_<uuid>/0f21e468-f125-4061-9b25-8df6c491141b,
...
Suppressed: java.nio.file.NoSuchFileException:
/tmp/junit12972024642090864369/<jobid>/taskowned/job_<jobid>_tm_<uuid>/0f21e468-f125-4061-9b25-8df6c491141b
{code}
https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=79018
https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=79008
https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=79000
https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=78965
It is not limited to Azure, the same mode shows up on release-1.20 push
runs on GitHub Actions:
https://github.com/apache/flink/actions/runs/33842483755
https://github.com/apache/flink/actions/runs/33783331299
release-2.2 and release-2.3 carry the same test without the fix, so they
are exposed as well.
> UnalignedCheckpointRescaleWithMixedExchangesITCase is unstable: @TempDir
> cleanup races with post-cancel async checkpoint
> ------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-40446
> URL: https://issues.apache.org/jira/browse/FLINK-40446
> Project: Flink
> Issue Type: Bug
> Reporter: Rui Fan
> Assignee: Rui Fan
> Priority: Major
> Labels: pull-request-available
>
> UnalignedCheckpointRescaleWithMixedExchangesITCase.testRescaleFromUnalignedCheckpoint
> is occasionally unstable. The test body passes, but teardown fails:
> {code:java}
> java.io.IOException: Failed to delete temp directory /tmp/junit-...
> Suppressed: java.nio.file.DirectoryNotEmptyException: /tmp/junit-...{code}
> Root cause: the test ends each job with `jobClient.cancel().get()`. cancel()
> is
> asynchronous, so .get() only means cancellation was acknowledged, not that
> the job
> reached a terminal state. The test then returns and JUnit deletes the
> @TempDir (which
> also serves as the checkpoint dir) while the source subtasks' async snapshot
> of the
> just-completed checkpoint is still writing SourceReaderState into the
> file-merging
> "taskowned" directory. The post-order delete walk then hits
> DirectoryNotEmptyException.
> It only reproduces with file-merging enabled
> (execution.checkpointing.file-merging.enabled
> = true), which is why it shows up on a fraction of runs.
> Fix: wait for the job to reach the CANCELED terminal state before proceeding,
> instead of
> relying on cancel().get() alone. (CommonTestUtils.terminateJob now waits for
> CANCELED,
> and the UC rescale ITCases go through it.)
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
