saintstack opened a new pull request #742: 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/742 … of XXXX, because it is not online on us" Fix is moving a line in HRegion to close a gap that would allow an old edit to sneak into the backing SequenceIdAccounting datastructure. The below line gets moved up before wal.startCacheFlush: - flushOpSeqId = getNextSequenceId(wal); We change the LogRoller constructor to remove redundant param. It will also call the temporary,ugly purge method (To be removed if we've found fix...). Because of constructor chagne, need to change also TestFailedAppendAndSync, TestWALLockup, TestAsyncFSWAL & WALPerformanceEvaluation.java We add a purge method called when something is wrong to AbstractFSWAL.java Need to be removed when 'fix' is in place. SequenceIdAccounting gets a purge method called when we find lagging records.
---------------------------------------------------------------- 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
