lianetm commented on PR #20254: URL: https://github.com/apache/kafka/pull/20254#issuecomment-3136300976
Hey all, interesting. I don't have too much context but it does look fishy when we `tryAppendForSplit` and always allow a first record (no matter the size it seems), here https://github.com/apache/kafka/blob/4c1b79851f6858f42ca0eab33afa88af1c531c1f/clients/src/main/java/org/apache/kafka/common/record/MemoryRecordsBuilder.java#L856 Definitely needs more investigation, maybe start by adding a test to check the retry behaviour around `splitAndReenqueue` for a large batch with a single record? I cannot find any test for it (here it's all about batches that can indeed be split https://github.com/apache/kafka/blob/4c1b79851f6858f42ca0eab33afa88af1c531c1f/clients/src/test/java/org/apache/kafka/clients/producer/internals/RecordAccumulatorTest.java#L1016 ) Hope it helps, I will stay in the loop. -- 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: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
