hbase的HMaster不能启动

在没有关闭hbase的情况下，集群电脑重新启动，导致HMaster不能启动，查看日志，有如下错误：
2014-01-20 16:55:24,217 INFO org.apache.hadoop.hbase.zookeeper.RecoverableZooKeeper: Node /hbase/master already exists and this is not a retry2014-01-20 16:55:24,219 INFO org.apache.hadoop.hbase.master.ActiveMasterManager: Adding ZNode for /hbase/backup-masters/RDCMaster.cluster,60000,1390208122540 in backup master directory2014-01-20 16:55:24,246 INFO org.apache.hadoop.hbase.master.ActiveMasterManager: Current master has this master's address, RDCMaster.cluster,60000,1390207622474; master was restarted? Deleting node.2014-01-20 16:55:24,247 DEBUG org.apache.hadoop.hbase.master.ActiveMasterManager: No master available. Notifying waiting threads2014-01-20 16:55:24,247 INFO org.apache.hadoop.hbase.master.HMaster: Cluster went down before this master became active2014-01-20 16:55:24,247 DEBUG org.apache.hadoop.hbase.master.HMaster: Stopping service threads2014-01-20 16:55:24,247 INFO org.apache.hadoop.ipc.HBaseServer: Stopping server on 600002014-01-20 16:55:24,247 INFO org.apache.hadoop.ipc.HBaseServer: IPC Server handler 0 on 60000: exiting2014-01-20 16:55:24,248 INFO org.apache.hadoop.ipc.HBaseServer: IPC Server handler 1 on 60000: exiting2014-01-20 16:55:24,248 INFO org.apache.hadoop.ipc.HBaseServer: IPC Server handler 2 on 60000: exiting2014-01-20 16:55:24,248 INFO org.apache.hadoop.ipc.HBaseServer: IPC Server handler 3 on 60000: exiting2014-01-20 16:55:24,248 INFO org.apache.hadoop.ipc.HBaseServer: IPC Server handler 4 on 60000: exiting2014-01-20 16:55:24,248 INFO org.apache.hadoop.ipc.HBaseServer: REPL IPC Server handler 0 on 60000: exiting2014-01-20 16:55:24,248 INFO org.apache.hadoop.ipc.HBaseServer: IPC Server handler 5 on 60000: exiting2014-01-20 16:55:24,248 INFO org.apache.hadoop.ipc.HBaseServer: REPL IPC Server handler 1 on 60000: exiting2014-01-20 16:55:24,248 INFO org.apache.hadoop.ipc.HBaseServer: REPL IPC Server handler 2 on 60000: exiting

问题在于HMaster的实例依然保存在hdfs上。

我配置的hbase-site.xml里hbase.rootdir为hdfs://RDCMaster.cluster:9000/hbase。

解决办法：

bin/hadoop dfs -rmr /hbase/hbase.id

重启集群电脑，接着启动hadoop, hbase，jps查看，HMaster已经正常启动！