[
https://issues.apache.org/jira/browse/FLINK-4510?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15545579#comment-15545579
]
ASF GitHub Bot commented on FLINK-4510:
---------------------------------------
Github user StefanRRichter commented on the issue:
https://github.com/apache/flink/pull/2453
Thanks for the nice contribution.
I see one minor detail that could cause problems. Even if we do not
register the `CheckpointCoordinator` as `JobStatusListener`, it might still be
possible that `triggerCheckpoint(long timestamp, CheckpointProperties props)`
gets called. In this case it can happen that a periodic trigger is registered.
You could introduce another check on the interval to make sure that no periodic
trigger can be registered when the CheckpointCoordinator is disabled.
Besides that, the changes look good to me.
> Always create CheckpointCoordinator
> -----------------------------------
>
> Key: FLINK-4510
> URL: https://issues.apache.org/jira/browse/FLINK-4510
> Project: Flink
> Issue Type: Sub-task
> Components: State Backends, Checkpointing
> Reporter: Ufuk Celebi
> Assignee: Jark Wu
>
> The checkpoint coordinator is only created if a checkpointing interval is
> configured. This means that no savepoints can be triggered if there is no
> checkpointing interval specified.
> Instead we should always create it and allow an interval of 0 for disabled
> periodic checkpoints.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
