jolshan commented on pull request #9944: URL: https://github.com/apache/kafka/pull/9944#issuecomment-828663698
Thanks @junrao for taking another look > A lot of the complexity is the additional logic for propagating unresolved partitions from FetchRequest to FetchSession and the maintenance of unresolved partitions within FetchSession. I agree. This has been in the back of my mind for a while, specifically whether all the changes in FetchSession are necessary for such cases. So thanks for bringing this up. I think the only thing I was really concerned about was during a roll to upgrade/the new topic case you mentioned. But even using the current approach, I wasn't sure if the fetch session stuff was really helping. I think my biggest confusion comes from when the client will refresh metadata. Will returning a top level error guarantee a refresh? (vs given an unknown topic ID response?) I think that the top level approach will likely be better. -- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
