[
https://issues.apache.org/jira/browse/FLINK-13497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16896752#comment-16896752
]
vinoyang commented on FLINK-13497:
----------------------------------
Currently, the {{CheckpointFailureManager}} choose a simple counting mechanism
to fail the job, so there is the possibility like [~till.rohrmann] said. It is
also another issue(FLINK-12514) to track a better counting mechanism.
The solution proposed by [~yunta] may fix this issue temporarily. But it may
cause another risk. The {{stopCheckpointScheduler}} also called
{{CheckpointFailureManager#handleCheckpointException}}. It will make counting
more complex in the future.
Maybe we need to call a pure method just fails all pending checkpoints when
failing the job?
> Checkpoints can complete after CheckpointFailureManager fails job
> -----------------------------------------------------------------
>
> Key: FLINK-13497
> URL: https://issues.apache.org/jira/browse/FLINK-13497
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Checkpointing
> Affects Versions: 1.9.0, 1.10.0
> Reporter: Till Rohrmann
> Priority: Critical
> Fix For: 1.9.0
>
>
> I think that we introduced with FLINK-12364 an inconsistency wrt to job
> termination a checkpointing. In FLINK-9900 it was discovered that checkpoints
> can complete even after the {{CheckpointFailureManager}} decided to fail a
> job. I think the expected behaviour should be that we fail all pending
> checkpoints once the {{CheckpointFailureManager}} decides to fail the job.
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)
