Apache9 commented on issue #739: HBASE-23157 WAL unflushed seqId tracking may wrong when Durability.AS… URL: https://github.com/apache/hbase/pull/739#issuecomment-545721751 I think this could be a follow on. It is the same problem, but we need to discuss how to fix it. If we think the current solution, where we remove all the related records when closing a region, is fine, then I think we could add some sequence ids back to the unflushedSequenceIds in completeCacheFlush, if we found that the current sequence id in the unflushedSequenceIds is even smaller then the flushedSequenceId. As we will write a flush marker and wait for it to complete, so after flush we can make sure that there are no pending WAL entries for this region any more, so I think the solution above can work.
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
