sodonnel commented on PR #10109: URL: https://github.com/apache/ozone/pull/10109#issuecomment-4345047536
This problem is kind of getting difficult to see if we have solved all the cases or not. The root cause is poor interface design, and further inadequate locking. Fixing this is going to be difficult. The locking is only inadequate because the lock is inside the container object and the container object can be replaced by a new object against the same ID. What are all the scenarios that replace that container object? What is we instead updated the existing object in place under a lock? Would that solve all these edge cases more simple? It **sounds like** these problems all come from the disk balancer moving and replacing the container object, so if we fix just that, it may be a cleaner solution to this whole problem? -- 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]
