julianbergner opened a new pull request, #22233: URL: https://github.com/apache/kafka/pull/22233
Clarifies in the producer `acks` documentation that records acknowledged under `acks=1` may be lost not only on hard leader failure, but also during planned leader transitions - partition reassignment that removes the current leader, preferred-leader election, and controlled shutdown of the leader broker - when those records have not yet reached the high watermark before leadership moves. This is expected behavior under the existing `acks=1` durability contract, but it is operationally surprising because the loss can happen during planned administrative operations rather than only after a hard failure. The behavior was reproduced on Kafka 4.1.2 and confirmed as expected on the dev mailing list (see JIRA for the thread reference). JIRA: https://issues.apache.org/jira/browse/KAFKA-20554 -- 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]
