Hadoop数据目录迁移

@(Hadoop)

随着数据的不断导入和增大，原本集群部署的目录磁盘空间不足了，所以要把hadoop存储数据的位置迁移到另外一个巨大的磁盘上，另外的一个用意是将数据和程序分离开，以免互相影响。

以下是迁移过程和需要注意的一些地方：

动手之前先把集群停止，如果有hbase也一起停了，因为hbase的存储是依赖于hdfs的，如果没有停止就进行目录迁移hbase会出现错误。

修改配置文件

hadoop最重要的存储数据的配置在core-site.xml文件中设置，修改core-site.xml的hadoop.tmp.dir值为新磁盘的路径即可。

考虑到数据和程序的分离，决定将那些会不断增长的文件都迁移出去，包括：日志文件，pid目录，journal目录。

日志文件和pid目录在hadoop-env.sh中配置，export HADOOP_PID_DIR，HADOOP_LOG_DIR为对应磁盘路径即可。

journal目录在hdfs-site.xml中配置dfs.journalnode.edits.dir

同理，yarn和hbase的log和pid文件路径都可在*_env.sh文件中export设置

改完Hadoop的配置文件之后将其拷贝到hbase/conf目录下

hbase的日志文件和pid目录配置在hbase-daemon.sh的HBASE_PID_DIR，HBASE_LOG_DIR

spark日志文件的pid目录在spark-env.sh的SPARK_PID_DIR，SPARK_LOG_DIR

修改完之后拷贝配置文件到各个子节点。

并将原始数据目录、日志目录和pid目录移动至新磁盘中，重新启动集群，查看输出信息是否正确。

更新

hdfs-site.xml中更新的配置：

<property>

<name>dfs.name.dir</name>

<value>/data2/hadoop/hdfs/name</value>

</property>

<property>

<name>dfs.data.dir</name>

<value>/data2/hadoop/hdfs/data</value>

</property>

分别是存储hdfs元数据信息和数据的目录，如果没有配置则默认存储到hadoop.tmp.dir中。

格式化hdfs系统之后，hbase启动异常，HMaster自动退出。

日志信息：

2016-01-15 14:01:38,231 DEBUG [MASTER_SERVER_OPERATIONS-zx-hadoop-210-11:60000-4] master.DeadServer: Finished processing zx-hadoop-210-24,60020,1452828414814

2016-01-15 14:01:38,231 ERROR [MASTER_SERVER_OPERATIONS-zx-hadoop-210-11:60000-4] executor.EventHandler: Caught throwable while processing event M_SERVER_SHUTDOWN

java.io.IOException: failed log splitting for zx-hadoop-210-24,60020,1452828414814, will retry

        at org.apache.hadoop.hbase.master.handler.ServerShutdownHandler.resubmit(ServerShutdownHandler.java:322)

        at org.apache.hadoop.hbase.master.handler.ServerShutdownHandler.process(ServerShutdownHandler.java:202)

        at org.apache.hadoop.hbase.executor.EventHandler.run(EventHandler.java:128)

        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)

        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)

        at java.lang.Thread.run(Thread.java:745)

Caused by: java.io.IOException: error or interrupted while splitting logs in [hdfs://ns1/hbase/WALs/zx-hadoop-210-24,60020,1452828414814-splitting] Task = installed =

 1 done = 0 error = 0

        at org.apache.hadoop.hbase.master.SplitLogManager.splitLogDistributed(SplitLogManager.java:362)

        at org.apache.hadoop.hbase.master.MasterFileSystem.splitLog(MasterFileSystem.java:410)

        at org.apache.hadoop.hbase.master.MasterFileSystem.splitLog(MasterFileSystem.java:384)

        at org.apache.hadoop.hbase.master.MasterFileSystem.splitLog(MasterFileSystem.java:282)

        at org.apache.hadoop.hbase.master.handler.ServerShutdownHandler.process(ServerShutdownHandler.java:195)

        ... 4 more

2016-01-15 14:01:38,232 INFO  [master:zx-hadoop-210-11:60000-EventThread] zookeeper.ClientCnxn: EventThread shut down

2016-01-15 14:01:38,232 INFO  [master:zx-hadoop-210-11:60000.oldLogCleaner] zookeeper.ZooKeeper: Session: 0x25243ddd648000a closed

2016-01-15 14:01:38,232 DEBUG [MASTER_SERVER_OPERATIONS-zx-hadoop-210-11:60000-4] master.DeadServer: Finished processing zx-hadoop-210-22,60020,1452828414925

2016-01-15 14:01:38,233 ERROR [MASTER_SERVER_OPERATIONS-zx-hadoop-210-11:60000-4] executor.EventHandler: Caught throwable while processing event M_SERVER_SHUTDOWN

java.io.IOException: Server is stopped

        at org.apache.hadoop.hbase.master.handler.ServerShutdownHandler.process(ServerShutdownHandler.java:183)

        at org.apache.hadoop.hbase.executor.EventHandler.run(EventHandler.java:128)

        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)

        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)

        at java.lang.Thread.run(Thread.java:745)

2016-01-15 14:01:38,338 DEBUG [master:zx-hadoop-210-11:60000] catalog.CatalogTracker: Stopping catalog tracker org.apache.hadoop.hbase.catalog.CatalogTracker@6c4b58f0

2016-01-15 14:01:38,338 INFO  [master:zx-hadoop-210-11:60000] client.HConnectionManager$HConnectionImplementation: Closing zookeeper sessionid=0x15243ddd6340004

2016-01-15 14:01:38,343 INFO  [master:zx-hadoop-210-11:60000] zookeeper.ZooKeeper: Session: 0x15243ddd6340004 closed

2016-01-15 14:01:38,343 INFO  [master:zx-hadoop-210-11:60000-EventThread] zookeeper.ClientCnxn: EventThread shut down

2016-01-15 14:01:38,343 INFO  [zx-hadoop-210-11,60000,1452837685871.splitLogManagerTimeoutMonitor] master.SplitLogManager$TimeoutMonitor: zx-hadoop-210-11,60000,14528

37685871.splitLogManagerTimeoutMonitor exiting

2016-01-15 14:01:38,347 INFO  [master:zx-hadoop-210-11:60000] zookeeper.ZooKeeper: Session: 0x35243ddd73b0001 closed

2016-01-15 14:01:38,347 INFO  [main-EventThread] zookeeper.ClientCnxn: EventThread shut down

2016-01-15 14:01:38,347 INFO  [master:zx-hadoop-210-11:60000] master.HMaster: HMaster main thread exiting

2016-01-15 14:01:38,350 ERROR [main] master.HMasterCommandLine: Master exiting

java.lang.RuntimeException: HMaster Aborted

        at org.apache.hadoop.hbase.master.HMasterCommandLine.startMaster(HMasterCommandLine.java:192)

        at org.apache.hadoop.hbase.master.HMasterCommandLine.run(HMasterCommandLine.java:134)

        at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:70)

        at org.apache.hadoop.hbase.util.ServerCommandLine.doMain(ServerCommandLine.java:126)

        at org.apache.hadoop.hbase.master.HMaster.main(HMaster.java:2785)

Fri Jan 15 14:15:02 CST 2016 Starting master on zx-hadoop-210-11

解决方法

1.切换到zookeeper的bin目录

2.执行$sh zkCli.sh

ls /

rmr /hbase

quit

重启hbase。

作者：@小黑