Apache9 commented on issue #739: HBASE-23181 Blocked WAL archive: "LogRoller: Failed to schedule flush of XXXX, because it is not online on us" URL: https://github.com/apache/hbase/pull/739#issuecomment-545733881 > What would be a follow on sir? > > Diff between 23181 and this is a sync wal. The patch on there at moment is bad I realize now. Like I proposed above, we could do something in the completeCacheFlush to prevent the flushed sequence id go backwards. And this does not lead to data loss, only a warning log on master side which makes people a bit confusing, so I think it is fine to address it in a follow on issue. Issuing a wal sync while holding the writeLock of updateLock is not good, and we have been doing a lot of works to avoid this... I've already changed the title to HBASE-23181, if we all agree that this is fine, let's merge this, and then let's use Lijin Bin's new UT to address the flushed sequence id problem using HBASE-23157. 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
