Dongnuo Lyu created KAFKA-21583:
-----------------------------------

             Summary: Return INVALID_COMMIT_OFFSET_SIZE from OffsetCommit and 
TxnOffsetCommit when the offsets exceed the maximum batch size
                 Key: KAFKA-21583
                 URL: https://issues.apache.org/jira/browse/KAFKA-21583
             Project: Kafka
          Issue Type: Bug
            Reporter: Dongnuo Lyu
            Assignee: Dongnuo Lyu


OffsetCommit and TxnOffsetCommit write one record for each committed partition 
in one atomic batch. When the batch exceeds max.message.bytes of 
__consumer_offsets, the old group coordinator returned 
INVALID_COMMIT_OFFSET_SIZE. The new group coordinator returns 
UNKNOWN_SERVER_ERROR, because CoordinatorOperationExceptionHelper maps 
MESSAGE_TOO_LARGE and RECORD_LIST_TOO_LARGE to it for every API.

The fix is to map MESSAGE_TOO_LARGE and RECORD_LIST_TOO_LARGE to 
INVALID_COMMIT_OFFSET_SIZE in GroupCoordinatorService.commitOffsets and 
commitTransactionalOffsets. The protocol stays unchanged.



--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

Reply via email to