allthingssecurity commented on PR #30132:
URL: https://github.com/apache/camel/pull/30132#issuecomment-6904650428

   Thanks for the second look. Pushed 18215af.
   
   **Finding 1 (batching processor):** I checked it and the batching path is 
not affected, so I left `KafkaRecordBatchingProcessor` as it is. Its break 
calls the no-argument `commitManager.commit()`, and that call never reaches 
Kafka when `batching=true`:
   
   - `KafkaConfiguration.isAutoCommitEnable()` returns `offsetRepository == 
null && !batching && autoCommitEnable`, so it is always false for a batching 
consumer, whatever `autoCommitEnable` is set to.
   - `SyncCommitManager.commit()` only calls `consumer.commitSync()` when 
`isAutoCommitEnable()` is true, and `AsyncCommitManager.commit()` the same for 
`commitAsync()`. `NoopCommitManager.commit()` and 
`CommitToOffsetManager.commit()` do nothing.
   
   So there is no `commitSync` in the batching break that can time out, and no 
`KafkaException` can leave `processException` to reach `startPolling`. To be 
sure I ran a throwaway copy of 
`KafkaConsumerBreakOnFirstErrorCommitTimeoutTest` with 
`batching=true&maxPollRecords=5` (manual sync commits, `autoCommitEnable` true 
and false, `pollOnError` `ERROR_HANDLER`, `DISCARD`, `RETRY`, `RECONNECT`), 
counting the calls of `commitSync()`: zero calls in the break, and the failed 
batch was polled again after the reconnect in every case. I did not keep it as 
a test since it pins a no-op. The PR description now says this more precisely.
   
   **Finding 2 (stale comment):** fixed. The comment above the commit of the 
break in `KafkaRecordStreamingProcessor` now says the commit can fail or time 
out, and the consumer still breaks out and polls again from the last committed 
offset.
   
   Comment-only change in code, so no new test. 
`KafkaConsumerBreakOnFirstErrorCommitTimeoutTest` passes (5 of 5) and the 
camel-kafka unit tests pass (349 run, 0 failures, 1 skipped).
   
   _Claude Code on behalf of allthingssecurity_
   


-- 
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