ArafatKhan2198 commented on PR #11934: URL: https://github.com/apache/ozone/pull/11934#issuecomment-6737993410
Thanks @devmadhuu for the review, and for catching this. You're right that retrying the whole batch isn't safe for ContainerKeyMapperTask. handlePutOMKeyEvent increments the global container count in SQL before the RocksDB mappings are committed. If the task throws partway through a batch, the retry can count the same containers again, and nothing would request a rebuild. My description assumed a thrown task had saved nothing, and that's only true for FileSizeCountTask. We also haven't found any real condition that makes these tasks throw. Given that, and since a correct fix means making those partial writes safe to repeat, which is more than this case justifies, I'm closing this PR and marking [HDDS-17040](https://issues.apache.org/jira/browse/HDDS-17040) as Won't Fix. If someone later finds a real trigger, we can reopen it with a safer approach, such as requesting a rebuild instead of retrying. Thanks again! -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
