[
https://issues.apache.org/jira/browse/FLINK-22268?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17322065#comment-17322065
]
Till Rohrmann commented on FLINK-22268:
---------------------------------------
I think the problem of this test is that the test only waits on the
{{JobStatus.RUNNING}} and then assumes that a synchronous savepoint succeeds.
However, the individual tasks of the job could still be in a different state
than {{ExecutionState.RUNNING}} and, hence, the savepoint can fail. Instead of
waiting on {{JobStatus.RUNNING}} we should check that all tasks are {{RUNNING}}.
cc [~dwysakowicz] because you seem to work on hardening these tests. It might
make sense to group all the {{JobMasterStopWithSavepointITCase}} test
instabilities since there exist quite some of them.
> JobMasterStopWithSavepointITCase.testRestartCheckpointCoordinatorIfStopWithSavepointFails
> fail because of "Not all required tasks are currently running."
> ---------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: FLINK-22268
> URL: https://issues.apache.org/jira/browse/FLINK-22268
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Checkpointing, Runtime / Coordination
> Affects Versions: 1.13.0
> Reporter: Guowei Ma
> Priority: Critical
> Labels: test-stability
>
> https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=16451&view=logs&j=5c8e7682-d68f-54d1-16a2-a09310218a49&t=f508e270-48d6-5f1e-3138-42a17e0714f0&l=3899
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
