[HBASE-2861] regionserver's logsyncer thread hangs on DFSClient$DFSOutputStream.waitForAckedSeqno

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Fix Version/s: 0.90.0
• Assignee: Hairong Kuang
• Reporter: Kannan Muthukkaruppan
• Created: 21/Jul/10 18:49
• Updated: 20/Nov/15 12:43
• Resolved: 01/Nov/10 06:23

Description

During loads into HBase, we are noticing that a RS is sometimes getting stuck.

The logSyncer thread:

      at org.apache.hadoop.hdfs.DFSClient$DFSOutputStream.waitForAckedSeqno(DFSClient.java:3367)
        - locked <0x00002aaac7fef748> (a java.util.LinkedList)
        at org.apache.hadoop.hdfs.DFSClient$DFSOutputStream.sync(DFSClient.java:3301)
        at org.apache.hadoop.fs.FSDataOutputStream.sync(FSDataOutputStream.java:97)
        at org.apache.hadoop.io.SequenceFile$Writer.syncFs(SequenceFile.java:944)
        at sun.reflect.GeneratedMethodAccessor3.invoke(Unknown Source)
        at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
        at java.lang.reflect.Method.invoke(Method.java:597)
        at org.apache.hadoop.hbase.regionserver.wal.SequenceFileLogWriter.sync(SequenceFileLogWriter.java:124)
        at org.apache.hadoop.hbase.regionserver.wal.HLog.hflush(HLog.java:949)

A lot of other threads are stuck on:

        at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:1925)
        at org.apache.hadoop.hbase.regionserver.wal.HLog$LogSyncer.addToSyncQueue(HLog.java:916)
        at org.apache.hadoop.hbase.regionserver.wal.HLog.sync(HLog.java:936)
        at org.apache.hadoop.hbase.regionserver.wal.HLog.append(HLog.java:828)
        at org.apache.hadoop.hbase.regionserver.HRegion.put(HRegion.java:1657)
        at org.apache.hadoop.hbase.regionserver.HRegion.put(HRegion.java:1425)
        at org.apache.hadoop.hbase.regionserver.HRegion.put(HRegion.java:1393)
        at org.apache.hadoop.hbase.regionserver.HRegionServer.put(HRegionServer.java:1665)
        at org.apache.hadoop.hbase.regionserver.HRegionServer.multiPut(HRegionServer.java:2326)

Subsequently, trying to disable the table, which in turn attempts to close the region(s), caused internalFlushCache() also to get stuck here:

      at java.util.concurrent.locks.AbstractQueuedSynchronizer.acquire(AbstractQueuedSynchronizer.java:1114)
        at java.util.concurrent.locks.ReentrantReadWriteLock$WriteLock.lock(ReentrantReadWriteLock.java:807)
        at org.apache.hadoop.hbase.regionserver.HRegion.internalFlushcache(HRegion.java:974)
        at org.apache.hadoop.hbase.regionserver.HRegion.close(HRegion.java:511)
        - locked <0x00002aaab76af670> (a java.lang.Object)
        at org.apache.hadoop.hbase.regionserver.HRegion.close(HRegion.java:463)
        at org.apache.hadoop.hbase.regionserver.HRegionServer.closeRegion(HRegionServer.java:1468)
        at org.apache.hadoop.hbase.regionserver.HRegionServer$Worker.run(HRegionServer.java:1329)

I'll attach the full jstack trace soon.

Issue Links

• is related to: HDFS-724 Pipeline close hangs if one of the datanode is not responsive.