C0urante commented on pull request #11196:
URL: https://github.com/apache/kafka/pull/11196#issuecomment-1068608386


   @xjrk58 have you observed this bug in production or verified with a unit 
test? The `WorkerSinkTask` class has undergone some modifications since this PR 
was made, but based on 
https://github.com/apache/kafka/blob/79d97bd29d059e8ba8ee7726b49d76e03e281059/connect/runtime/src/main/java/org/apache/kafka/connect/runtime/WorkerSinkTask.java#L708-L709
 (which was present when this PR was filed), it looks like the message batch 
was getting cleared any time partitions were revoked from the consumer, which 
should happen if the poll timeout elapses and the consumer is kicked out of the 
group.
   
   We have similar logic in the class today, but which is more granular about 
exactly which messages are removed from the batch (only the ones that 
originated from revoked partitions are removed): 
https://github.com/apache/kafka/blob/620f1d88d80fdf8150bd0b75be307bc4a2d3a0ea/connect/runtime/src/main/java/org/apache/kafka/connect/runtime/WorkerSinkTask.java#L773-L774


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


Reply via email to