Nicklee007 commented on PR #3773: URL: https://github.com/apache/bookkeeper/pull/3773#issuecomment-1445915568
> I have one concern about the sort. If we have multiple replication workers, and each replication worker sorts the ledgers before getting the target ledger to replicate. Would this PR introduce more Zookeeper try-lock operation conflicts? @hangc0276 we set a condition `if (depth == 3)` to avoid the conflicts. The zk under replicate ledger dir have 4 layers under /unserreplication/ledgers, like `/aaaa/bbbb/cccc/dddd`. So, in the PR only sorted the first three layers `/aaaa/bbbb/cccc` to keep we can replicate the fresh ledger at first as far as possible, and the last layer `/dddd` each worker shuffled znode to avoid workers always try-lock the same znode and operation conflicts. -- 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]
