-
Bug
-
Resolution: Done
-
Major
-
2.6.2.Final, 2.7.0.Beta1
Bug report
For bug reports, provide this information, please:
What Debezium connector do you use and what version?
debezium-connector-jdbc latest
What is the connector configuration?
Any
What is the captured database version and mode of depoyment?
Any
What behaviour do you expect?
I expect the close method in JdbcSinkConnectorTask to clear the topic partitions offsets when it it closed as it might get re-assigned with different set of topic partitions. The re-assignment can happen due to re-balance. The task shouldn't be holding on to old topic partitions.
Other sink connectors which handle this:
What behaviour do you see?
If the topic partitions weren't cleared, it results in below error,
Ignoring invalid task provided offset topicXYZ-1 OffsetAndMetadata{offset=4966, leaderEpoch=null, metadata=''} -- partition not assigned, assignment=[topicXYZ-1] (org.apache.kafka.connect.runtime.WorkerSinkTask:434)
Do you see the same behaviour using the latest relesead Debezium version?
Yes
Do you have the connector logs, ideally from start till finish?
Added above
How to reproduce the issue using our tutorial deployment?
Reproduced using unit-test by asserting the contents of sinkTask.offsets.
- links to
-
RHEA-2024:139598 Red Hat build of Debezium 2.5.4 release