Umeshkumar9414 commented on PR #7075: URL: https://github.com/apache/hbase/pull/7075#issuecomment-3061080761
> As I asked above, if we do not change the mechanism that we will only schedule a new split worker while the previous one is complete or the region server is dead, what is the advantage of using different file names for split workers? @Apache9 In the wal splitting using recovered edits output sink, worker first writes the temporary edit files for each regions and then rename it to actual recovered edits. If a worker find a existing temporary edit file it deletes the temporary edit file and creates a new one. In our case where a zombie worker and a new worker are splitting the wal at the same time, we don't have any issue if new worker deletes the temporary files but zombie worker can also delete the temporary files being used by new worker causing new worker to fail with FileNotFoundException. If zombie worker is slow then a new worker can fail multiple times, blocking the ServerCrashProcedure for extended period of time, thus reducing avaialability. Using differenet name ensures that both are not interferring with each other. -- 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: issues-unsubscr...@hbase.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
