[
https://issues.apache.org/jira/browse/FLINK-30539?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17653578#comment-17653578
]
Matthias Pohl edited comment on FLINK-30539 at 1/2/23 12:07 PM:
----------------------------------------------------------------
Thanks for sharing your opinion [~gaborgsomogyi].
{quote}By eliminating timeout you mean we should depend on CI timeout?{quote}
Yes, because the CI timeout (or actually the watcher process) would trigger a
thread dump that would reveal where the test is hanging. That said, removing
the timeouts is not a proposal for fixing the test instability but to comply to
our [coding
conventions|https://flink.apache.org/contributing/code-style-and-quality-common.html#avoid-timeouts-in-junit-tests]
which would help us dig into what's going on.
was (Author: mapohl):
Thanks for sharing your opinion [~gaborgsomogyi].
> By eliminating timeout you mean we should depend on CI timeout?
Yes, because the CI timeout (or actually the watcher process) would trigger a
thread dump that would reveal where the test is hanging. That said, removing
the timeouts is not a proposal for fixing the test instability but to comply to
our [coding
conventions|https://flink.apache.org/contributing/code-style-and-quality-common.html#avoid-timeouts-in-junit-tests]
which would help us dig into what's going on.
> YARNSessionCapacitySchedulerITCase.testDetachedPerJobYarnCluster and
> testDetachedPerJobYarnClusterWithStreamingJob timing out
> -----------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-30539
> URL: https://issues.apache.org/jira/browse/FLINK-30539
> Project: Flink
> Issue Type: Bug
> Components: Deployment / YARN
> Affects Versions: 1.16.0
> Reporter: Matthias Pohl
> Priority: Major
> Labels: test-stability
>
> https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=44337&view=logs&j=298e20ef-7951-5965-0e79-ea664ddc435e&t=d4c90338-c843-57b0-3232-10ae74f00347&l=32023
> Both tests failed because they are running into the 60s timeout that was
> defined for each of them. We should get rid of the timeout to access the
> thread dump. It might be related to FLINK-24169
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
