F64116045 commented on PR #11199: URL: https://github.com/apache/ozone/pull/11199#issuecomment-5780914971
> I think we should try to avoid them, otherwise we end up with an explosion of states. The original idea was that a container can only be corrected for one state at a time to keep things simple RM should correct one issue at a time makes sense to me. > hence it should not be under-replicated + something else. Once it comes out of under replicated it can go to something els But this would represent similar cases differently. For example: ```text Actual issues: UNDER_REPLICATED + DATA_CHECKSUM_MISMATCH Recorded state: UNDER_REPLICATED ``` But the existing `UNHEALTHY_UNDER_REPLICATED` case, the handler [records both conditions but only queues the under-replication repair](https://github.com/apache/ozone/blob/5a73b2688f/hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/container/replication/health/RatisUnhealthyReplicationCheckHandler.java#L71-L86): ```text Actual issues: UNHEALTHY + UNDER_REPLICATED Recorded state: UNHEALTHY_UNDER_REPLICATED ``` We would then have to know which conditions are combined, which are hidden until another issue is resolved, and why. I think maybe we need a consistent rule for that. -- 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]
