guozhangwang commented on PR #11424: URL: https://github.com/apache/kafka/pull/11424#issuecomment-1460510199
Thanks for keep looking into this issue, and continue to try getting it merged. After chatting with you offline I thought a bit more around the logic of pause / resume. And I think that besides the additional logic of: 1) pause all non-empty partitions when the global bytes threshold is exceeded; 2) resume all paused partitions when below the global bytes threshold. We also do the following: 3) modify the single partition resume logic (a.k.a. https://github.com/apache/kafka/pull/11424/files#diff-a76674468cda8772230fb8411717cf9068b1a363a792f32c602fb2ec5ba9efd7R722) that if the partition buffer becomes empty after the last record is retrieved, resume it if it was paused. I.e. we should remove the TODO since it would not be removed, and also not condition on `maxBufferedSize != -1`. 4) deprecated the single partition pause logic (a.k.a. https://github.com/apache/kafka/pull/11424/files#diff-a76674468cda8772230fb8411717cf9068b1a363a792f32c602fb2ec5ba9efd7R977) just like what you did previously still, since we would still remove it in the future. cc @mjsax @vvcephei who also looked into this. -- 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
