ytong01 opened a new issue, #4680: URL: https://github.com/apache/bookkeeper/issues/4680
**BUG REPORT** ***Describe the bug*** While client reads entry(s)，bookie will send read request to node in sequence, and WriteSet is recycled when we receive a successfully response, but SequenceReadRequest will try to registerSlowBookie according writeSet and nextReplicaIndexToReadFrom later, unfortunately the WriteSet is already recycled and a IndexOutOfBoundsException is raised unexpected. ***To Reproduce*** Steps to reproduce the behavior: Assume we declare a configuration like this, ensembleSize 4, writeQuorumSize 2, ackQuorumSize 2 1. Client call LedgerHandle#readEntry, and assume we will read (0,1) in ensemble 2. if bookieIndex 0 read failed, we continue read from 1 and read successfully 3. Invoke SequenceReadRequest#complete and recycle WriteSet 4. Try to registerSlowBookies subsequence but the WriteSet is recycled, error is occurs. Hope I have make it sense and look forward to your reply soon, thanks. -- 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]
