ivandika3 commented on PR #7459: URL: https://github.com/apache/ozone/pull/7459#issuecomment-2488156594
@sodonnel Thanks for taking a look at this. > All that logic is there for some reason, so just removing it seems risky Yes, I'm not really well-versed in the design decisions for the container move in https://issues.apache.org/jira/browse/HDDS-5253. Currently, I'm looking at possible reason why this particular code is introduced and what is the possible impact of removing this code. I'm open for suggestions from any of the community members are familiar with this. > What version are you currently running? Our main cluster version is based on 1.2.1, but we have backported a significant amount of features and fixes. We are planning to upgrade our main cluster to 1.4.1 very soon. However, even in 1.4.1 this particular issue might still happen (albeit very rare). Therefore, I believe we need to address this since LegacyReplicationManager is still technically still supported. > Can you switch off the LegacyRM and use the new one? Yes, due to this bug we are now strongly considering to use the new ReplicationManager that do not persistently keep track of the inflight move. Can I verify whether there are some known regressions in moving from LegacyReplicationManager to ReplicationManager? I remember asking this quite a while ago whether we can just upgrade the implementation from LegacyReplicationManager to ReplicationManager without regressions and seems there might not be any regressions. -- 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]
