[
https://issues.apache.org/jira/browse/FLINK-4753?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15571319#comment-15571319
]
ASF GitHub Bot commented on FLINK-4753:
---------------------------------------
Github user tzulitai commented on the issue:
https://github.com/apache/flink/pull/2604
@rmetzger I think I just wanted to make sure that the
`KafkaTopicPartitionState<KPH>[]` state in the fetcher wasn't concurrently
accessed, but actually on a second look there shouldn't be any problems with
this (not really sure what was going through my head back then :p ). And you're
right, this doesn't add any additional correctness for the periodic committer.
Will close this and the ticket. Thanks for taking a look anyways!
> Kafka 0.8 connector's Periodic Offset Committer should synchronize on
> checkpoint lock
> -------------------------------------------------------------------------------------
>
> Key: FLINK-4753
> URL: https://issues.apache.org/jira/browse/FLINK-4753
> Project: Flink
> Issue Type: Bug
> Components: Kafka Connector
> Reporter: Tzu-Li (Gordon) Tai
> Assignee: Tzu-Li (Gordon) Tai
> Fix For: 1.2.0
>
>
> The {{PeriodicOffsetCommitter}} doesn't synchronize on the checkpoint lock
> when creating a copy of the partition offsets in the current state.
> I propose to pass the checkpoint lock into{{PeriodicOffsetCommitter}}, and
> let {{PeriodicOffsetCommitter}} use {{Kafka08Fetcher#snapshotCurrentState()}}
> to make the current state copy.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
