[HBASE-8246] Backport HBASE-6318 to 0.94 where SplitLogWorker exits due to ConcurrentModificationException

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.94.6
• Fix Version/s: 0.94.7
• Component/s: regionserver
• Hadoop Flags: Reviewed
• Assignee: Ted Yu
• Reporter: Jeffrey Zhong
• Created: 02/Apr/13 21:03
• Updated: 27/Apr/13 15:55
• Resolved: 03/Apr/13 13:48

Description

Today we found the following error in our tests. Later I found we already fixed the issue in trunk. I think we should backpor the fix because the consequence of the issue is high and the fix isn't complicated.

2013-04-01 21:23:21,864 INFO org.apache.hadoop.hbase.regionserver.SplitLogWorker: worker ip-10-143-160-121.ec2.internal,60020,1364849529986 done with task /hbase/splitlog/hdfs%3A%2F%2Fip-10-137-16-140.ec2.internal%3A8020%2Fapps%2Fhbase%2Fdata%2F.logs%2Fip-10-137-20-188.ec2.internal%2C60020%2C1364849530779-splitting%2Fip-10-137-20-188.ec2.internal%252C60020%252C1364849530779.1364865556657 in 67129ms
2013-04-01 21:23:21,864 ERROR org.apache.hadoop.hbase.regionserver.SplitLogWorker: unexpected error
java.util.ConcurrentModificationException
        at java.util.TreeMap$PrivateEntryIterator.nextEntry(TreeMap.java:1100)
        at java.util.TreeMap$ValueIterator.next(TreeMap.java:1145)
        at org.apache.hadoop.hbase.regionserver.wal.HLogSplitter$OutputSink.closeLogWriters(HLogSplitter.java:1279)
        at org.apache.hadoop.hbase.regionserver.wal.HLogSplitter$OutputSink.finishWritingAndClose(HLogSplitter.java:1170)
        at org.apache.hadoop.hbase.regionserver.wal.HLogSplitter.splitLogFile(HLogSplitter.java:475)
        at org.apache.hadoop.hbase.regionserver.wal.HLogSplitter.splitLogFile(HLogSplitter.java:403)
        at org.apache.hadoop.hbase.regionserver.SplitLogWorker$1.exec(SplitLogWorker.java:111)
        at org.apache.hadoop.hbase.regionserver.SplitLogWorker.grabTask(SplitLogWorker.java:264)
        at org.apache.hadoop.hbase.regionserver.SplitLogWorker.taskLoop(SplitLogWorker.java:195)
        at org.apache.hadoop.hbase.regionserver.SplitLogWorker.run(SplitLogWorker.java:163)
        at java.lang.Thread.run(Thread.java:662)
2013-04-01 21:23:21,865 INFO org.apache.hadoop.hbase.regionserver.SplitLogWorker: SplitLogWorker ip-10-143-160-121.ec2.internal,60020,1364849529986 exiting

The impact of this issue is that SplitLogWorker exits so does the region server recovering mechanism of HBase. If any RS failed after all SplitLogWorkers in te cluster exit due to the issue, you'll see a hang log splitting job and the failed RS won't be recovered.