You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Yun Tang (JIRA)" <ji...@apache.org> on 2019/07/15 10:22:00 UTC
[jira] [Commented] (FLINK-13256) Periodical checkpointing is
stopped after failovers
[ https://issues.apache.org/jira/browse/FLINK-13256?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16885043#comment-16885043 ]
Yun Tang commented on FLINK-13256:
----------------------------------
After digging into the heap dump of jobmanager provided by [~zhuzh], I figured out the root cause. To be short, this is due to {{currentPeriodicTrigger}} in {{CheckpointCoordinator}} had been assigned to {{null}} during triggering checkpoint as there still existed a pending {{checkpoint-387}}. However, the pending checkpoints would then be aborted due to the region failover and no active {{currentPeriodicTrigger}} anymore.
> Periodical checkpointing is stopped after failovers
> ---------------------------------------------------
>
> Key: FLINK-13256
> URL: https://issues.apache.org/jira/browse/FLINK-13256
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Checkpointing
> Affects Versions: 1.9.0
> Reporter: Zhu Zhu
> Assignee: Yun Tang
> Priority: Blocker
> Fix For: 1.9.0
>
> Attachments: 15_15_20__07_15_2019.jpg, jm_no_cp_after_failover.log
>
>
> In this case, we observed that the job initially is triggering periodical checkpoints as expected.
> But after 2 region failovers, no checkpoint is triggered any more, even after all the tasks are RUNNING again.
> A sample log is attached along with the related topology desc pic.
> This case may not be reproduced every time.
>
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)