kamalcph commented on PR #14301: URL: https://github.com/apache/kafka/pull/14301#issuecomment-1697246090
> Okay, but I still do not understand, what is preventing the value in the log-start-checkpoint-offset from being stale as well? Or is the idea here that we will start with the log-start-*-offset which we knew was correct at some point in time and then we will contact tiered storage to try and figure out what later value it might have moved to? If so, does it matter which offset we start from as long as we have the logic for checking what is in tiered storage? yes, correct. The log-start-offset will be updated to the correct value eventually when the rlm-task updates the log-start-offset after becoming leader (patch needs to be opened). > If so, does it matter which offset we start from as long as we have the logic for checking what is in tiered storage? If the log-start-offset moves ahead, then currently remote-log-segments are deleted due to breach by log-start-offset. Once the rlm-task updates the log-start-offset after becoming leader, then those remote-log-segments won't be moved. But, the log-start-offset should increase monotonically, it should not go back and forth. -- 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
