sodonnel commented on pull request #1338: URL: https://github.com/apache/hadoop-ozone/pull/1338#issuecomment-698259315
Sorry for the slow reply on this. I have been caught up on some other things. > After a second thought, deleting the container record in SCM DB immediately while keep it in memory maybe a better and clean choice. So if there is stale container replica, it can be deleted based on in memory information. I think this is a good enough idea for now. If SCM is up for a very long time, perhaps in the future we will want to add a thread to clear all the in memory DELETED containers. One small concern is that if a container goes DELETED and then SCM is restarted soon after. Then a DN is restarted and reports a stale replica, it will just be seen as an unknown container. The default position there, is to log a warning. The config hdds.scm.unknown-container.action controls this. This is all an edge case - most of the time, all DNs should be up anyway. I left just one comment on a suggested refactor in the container report handler, when dealing with replicas from a DELETED container. Could you also add a test in TestContainerReportHander to check the logic around deleting a replica from a DELETED container? ---------------------------------------------------------------- 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 --------------------------------------------------------------------- To unsubscribe, e-mail: ozone-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: ozone-issues-h...@hadoop.apache.org
