frankvicky commented on PR #19269: URL: https://github.com/apache/kafka/pull/19269#issuecomment-2757028923
Hi @ableegoldman, > can you look into adding a test to make sure that the producer does still get reset/recreated properly if the producer is reset (eg transaction hits a timeout exception) but the StreamThread is not shutting down? I've investigated the related code and found writing a test for this specific scenario challenging. The main difficulty is that `TaskManager` doesn't control the timeout logic. While `TaskManager#handleLostAll` invokes `ActiveTaskCreator#reInitializeProducer`, this method only triggers when a `TaskMigratedException` occurs. The entire process is managed by `StreamThread`. The main pain point in writing this test lies in the deeply nested component structure: `StreamsProducer` is a member of `ActiveTaskCreator`, which is a member of `TaskManager`, which is ultimately controlled by `StreamThread`. When using mocks, this multi-layered nesting makes testing extremely complex. We typically mock outer components, but this makes it difficult to directly access and verify the reset behavior of the inner `StreamsProducer`. -- 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
