[
https://issues.apache.org/jira/browse/FLINK-36532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18127246#comment-18127246
]
Han You commented on FLINK-36532:
---------------------------------
I opened a PR here https://github.com/apache/flink-connector-kafka/pull/333
> Kafka group offsets are not committed on stop with savepoint
> ------------------------------------------------------------
>
> Key: FLINK-36532
> URL: https://issues.apache.org/jira/browse/FLINK-36532
> Project: Flink
> Issue Type: Bug
> Components: Connectors / Kafka
> Affects Versions: 1.18.1
> Reporter: LEONID ILYEVSKY
> Priority: Major
> Labels: pull-request-available
>
> In the pipeline with kafka source, with checkpointing enabled, on every
> checkpoint kafka source will commit the group offsets to the broker.
> However, when the job is stopped with savepoint, it does not commit those
> offsets.
> I verified this behavior in versions 1.16.1 and 1.18.1.
> In some situations, committing group offsets upon the job stop is very
> important. The best example is, when program code changes and the savepoint
> from the previous version is not compatible with the new code. If we cannot
> use the savepoint for the initial offsets, then we could use the group
> offsets, so it is critical to have them committed.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
