AndrewJSchofield opened a new pull request, #15479: URL: https://github.com/apache/kafka/pull/15479
PR https://github.com/apache/kafka/pull/13760 introduced a problem with KafkaAdmin.deleteRecords. If the request acted on a set of topic-partitions which spanned multiple leader brokers, the request for all of the topic-partitions were sent to all brokers. While this technically worked, it did mean that every broker handled its own topic-partitions and failed all of the ones that it didn't lead. This meant that every topic-partition was acted on, but at the cost of a lot of failed subrequests. The code was not paying attention to the mapping from node to topic-partition passed into `DeleteRecordsHandler.buildBatchedRequest`. This PR filters the subrequests for each node based on the mapping passed into that method. One of the existing unit tests actually codified the wrong behavior, so that has been fixed. Then a more complicated unit test that generates a mapping and then checks that it is correctly filtered has been added. ### Committer Checklist (excluded from commit message) - [ ] Verify design and implementation - [ ] Verify test coverage and CI build status - [ ] Verify documentation (including upgrade notes) -- 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