FATAL master.HMaster：意外状态：..无法将其转换为离线状态

我遇到了严重的 Hbase 崩溃问题。我使用 HBase 0.94.7 和一台主服务器和两台区域服务器。 HBase master 经常崩溃，我什至无法重新启动它。我的主日志如下：

DEBUG master.AssignmentManager: Handling transition=RS_ZK_REGION_CLOSED, server=master,60020,1374506461230, region=46c2333f401964bf877254be19c2cc8c
DEBUG handler.ClosedRegionHandler: Handling CLOSED event for 6423df864603aa6e8c45c726ab3ae62f
DEBUG master.AssignmentManager: Forcing OFFLINE; was=LogDetail,\x00\x00\x01\xE8\x00\x00\x01?\xF8\xB3\x8F\x17\xCE\xE2g\x84,1374498065657.6423df864603aa6e8c45c726ab3ae62f. state=CLOSED, ts=1374508769672, server=slave,60020,1374506460892
DEBUG zookeeper.ZKAssign: master:60000-0x14006f52f3f000e Creating (or updating) unassigned node for 6423df864603aa6e8c45c726ab3ae62f with OFFLINE state
FATAL master.HMaster: Unexpected state : LogDetail,\x00\x00\x01\xE8\x00\x00\x01?\xF6\xC17p&c\x8F\x14,1374498085655.c2f4143750eb1559a1dd92e937ea712d. state=PENDING_OPEN, ts=1374508769697, server=master,60020,1374506461230 .. Cannot transit it to OFFLINE.
java.lang.IllegalStateException: Unexpected state : LogDetail,\x00\x00\x01\xE8\x00\x00\x01?\xF6\xC17p&c\x8F\x14,1374498085655.c2f4143750eb1559a1dd92e937ea712d. state=PENDING_OPEN, ts=1374508769697, server=master,60020,1374506461230 .. Cannot transit it to OFFLINE.
    at org.apache.hadoop.hbase.master.AssignmentManager.setOfflineInZooKeeper(AssignmentManager.java:1879)
    at org.apache.hadoop.hbase.master.AssignmentManager.assign(AssignmentManager.java:1688)
    at org.apache.hadoop.hbase.master.AssignmentManager.assign(AssignmentManager.java:1424)
    at org.apache.hadoop.hbase.master.AssignmentManager.assign(AssignmentManager.java:1399)
    at org.apache.hadoop.hbase.master.AssignmentManager.assign(AssignmentManager.java:1394)
    at org.apache.hadoop.hbase.master.handler.ClosedRegionHandler.process(ClosedRegionHandler.java:105)
    at org.apache.hadoop.hbase.executor.EventHandler.run(EventHandler.java:175)
    at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
    at java.lang.Thread.run(Thread.java:662)
INFO master.HMaster: Aborting
DEBUG handler.ClosedRegionHandler: Handling CLOSED event for 0710b486dcb3d51465695b51db376255

....

DEBUG master.AssignmentManager: The znode of region LogDetail,\x00\x00\x01\xE8\x00\x00\x01?\xF6\xC17p&c\x8F\x14,1374498085655.c2f4143750eb1559a1dd92e937ea712d. has been deleted.
INFO master.AssignmentManager: The master has opened the region LogDetail,\x00\x00\x01\xE8\x00\x00\x01?\xF6\xC17p&c\x8F\x14,1374498085655.c2f4143750eb1559a1dd92e937ea712d. that was online on master,60020,1374506461230
DEBUG master.AssignmentManager: Handling transition=M_ZK_REGION_OFFLINE, server=master,60000,1374508461536, region=c9cfdd360c09b292412ba5ad88815e6f
DEBUG catalog.CatalogTracker: Stopping catalog tracker org.apache.hadoop.hbase.catalog.CatalogTracker@5c061cd2
INFO client.HConnectionManager$HConnectionImplementation: Closed zookeeper sessionid=0x14006f52f3f000f
INFO zookeeper.ZooKeeper: Session: 0x14006f52f3f000f closed
INFO zookeeper.ClientCnxn: EventThread shut down
INFO master.AssignmentManager$TimerUpdater: master,60000,1374508461536.timerUpdater exiting
INFO master.SplitLogManager$TimeoutMonitor: master,60000,1374508461536.splitLogManagerTimeoutMonitor exiting
INFO master.AssignmentManager$TimeoutMonitor: master,60000,1374508461536.timeoutMonitor exiting
INFO zookeeper.ZooKeeper: Session: 0x14006f52f3f000e closed
INFO zookeeper.ClientCnxn: EventThread shut down
INFO master.HMaster: HMaster main thread exiting
ERROR master.HMasterCommandLine: Failed to start master

我还在ZK日志中发现了一些不寻常的东西：

INFO org.apache.zookeeper.server.NIOServerCnxnFactory: Accepted socket connection from /master:37856
INFO org.apache.zookeeper.server.ZooKeeperServer: Client attempting to establish new session at /master:37856
INFO org.apache.zookeeper.server.ZooKeeperServer: Established session 0x140100dda0300e1 with negotiated timeout 180000 for client /master:37856
WARN org.apache.zookeeper.server.NIOServerCnxn: caught end of stream exception
EndOfStreamException: Unable to read additional data from client sessionid 0x140100dda0300e1, likely client has closed socket
        at org.apache.zookeeper.server.NIOServerCnxn.doIO(NIOServerCnxn.java:220)
        at org.apache.zookeeper.server.NIOServerCnxnFactory.run(NIOServerCnxnFactory.java:208)
        at java.lang.Thread.run(Thread.java:662)
INFO org.apache.zookeeper.server.NIOServerCnxn: Closed socket connection for client /master:37856 which had sessionid 0x140100dda0300e1

有人可以帮忙看看问题出在哪里吗？是否与未分配区域或类似的东西有关？我已经尝试过bin/hbase hbck -repair and bin/hbase hbck -fix，但这没有帮助。

Thanks

仔细检查了我的区域服务器的日志后，我得到了答案。

Cause

事实证明，区域服务器上没有很好地安装一个名为“SNAPPY”的用于压缩 hbase 表的库。我的所有表都是使用这种压缩算法创建的。当主节点尝试将区域平衡到区域服务器时，它失败了。最终主人流产了。

Solution

安装并配置 SNAPPY每个节点如下：

apt-get install libsnappy1
su hbase
mkdir /home/hbase/hbase-0.94.7/lib/native/Linux-amd64-64
ln -s /usr/lib/libsnappy.so.1.1.2 /home/hbase/hbase-0.94.7/lib/native/Linux-amd64-64/libsnappy.so
exit (-> root)
ln -s /usr/lib/libsnappy.so.1.1.2 /usr/lib64/libsnappy.so.1.1.2
ln -s /usr/lib/libsnappy.so.1.1.2 /usr/lib64/libsnappy.so.1
ln -s /usr/lib/libsnappy.so.1.1.2 /usr/lib64/libsnappy.so
ln -s /usr/lib/libsnappy.so.1 /usr/lib/libsnappy.so

现在一切都OK了！这些区域在区域服务器上非常平衡。