HBASE-多节点安装 UBUNTU 11

本文在hadoop-0.20.203.0  多节点安装后，在其上继续安装hbase-0.90.3

hadoop安装参考：
http://han-zoo.iteye.com/admin/blogs/1274926

---------------------------------------
总体步骤：在namenode上解压配置hbase->传送到其他NODE->回到namenode上启动hbase

详细步骤：

一、配置HBASE

1、在namenode 的DM1机器上下载hbase-0.90.3

2、解压在单独目录：/usr/hadoop/hbase-0.90.3

3、配置文件 / hbase-site.xml

/usr/hadoop/hbase-0.90.3/conf/ hbase-site.xml

注意： hdfs://dm1:9000/taisan 在HADOOP上需要建立

<configuration>

<property>
<name>hbase.rootdir</name>
<value> hdfs://dm1:9000/taisan </value> 
<description>hadoop dir</description>
</property>

<property>
<name>hbase.cluster.distributed</name>
<value>true</value>
</property>

<property>
<name>hbase.master</name>
<value>192.168.40.240:60000</value>
</property>

<property>
<name>hbase.zookeeper.quorum</name>
<value>192.168.10.147,192.168.40.242,192.168.40.239</value>  备注：另外三台的IP
</property>

</configuration>

4、配置文件 hbase-env.sh

/usr/hadoop/hbase-0.90.3/conf /hbase-env.sh

配置java环境如下：

export JAVA_HOME=/usr/java/jdk1.6.0_27/


5、配置文件regionservers
192.168.10.147
192.168.40.242
192.168.40.239

----------------------------------------------------------------------------------------------------------------
至此配置完毕，如果此时传送到其他机器，启动时会报错
ERROR: org.apache.hadoop.hbase.MasterNotRunningException: null

需要作如下操作：

复制hadoop/lib下的 commons-configuration-1.6.jar  到hbase/lib下
删除hbase/lib下的hadoop-core-0.20-append-r1056497.jar
复制hadoop/lib下的hadoop-core-0.20.203.0.jar 到hbase/lib下

操作记录：
cp /usr/hadoop/hadoop-0.20.203.0/lib/commons-configuration-1.6.jar /usr/hadoop/hbase-0.90.3/lib/.
chmod 715 /usr/hadoop/hbase-0.90.3/lib/commons-configuration-1.6.jar

rm /usr/hadoop/hbase-0.90.3/lib/hadoop-core-0.20-append-r1056497.jar
cp /usr/hadoop/hadoop-0.20.203.0/hadoop-core-0.20.203.0.jar /usr/hadoop/hbase-0.90.3/lib/.

chmod 715 /usr/hadoop/hbase-0.90.3/lib/hadoop-core-0.20.203.0.jar

----------------------------------------------------------------------------------------------------------------

6、传输到其他机器
scp -r hbase-0.90.3/ dm2:/usr/hadoop/.
scp -r hbase-0.90.3/ dm3:/usr/hadoop/.
scp -r hbase-0.90.3/ dm4:/usr/hadoop/.


7、启动hbase

hadoop@dm1:/usr/hadoop/hbase-0.90.3$ ./bin/start-hbase.sh
192.168.40.242: starting zookeeper, logging to /usr/hadoop/hbase-0.90.3/bin/../logs/hbase-hadoop-zookeeper-dm3.out
192.168.40.239: starting zookeeper, logging to /usr/hadoop/hbase-0.90.3/bin/../logs/hbase-hadoop-zookeeper-dm4.out
192.168.10.147: starting zookeeper, logging to /usr/hadoop/hbase-0.90.3/bin/../logs/hbase-hadoop-zookeeper-dm2.out
starting master, logging to /usr/hadoop/hbase-0.90.3/bin/../logs/hbase-hadoop-master-dm1.out
192.168.40.242: starting regionserver, logging to /usr/hadoop/hbase-0.90.3/bin/../logs/hbase-hadoop-regionserver-dm3.out
192.168.40.239: starting regionserver, logging to /usr/hadoop/hbase-0.90.3/bin/../logs/hbase-hadoop-regionserver-dm4.out
192.168.10.147: starting regionserver, logging to /usr/hadoop/hbase-0.90.3/bin/../logs/hbase-hadoop-regionserver-dm2.out

观察master：

hadoop@dm1:/usr/hadoop/hbase-0.90.3$ jps
10009 HMaster
5477 NameNode
5744 JobTracker
10255 Jps
5665 SecondaryNameNode


观察节点上：

hadoop@dm2:/usr/hadoop$ jps
5572 HQuorumPeer
5737 Jps
5690 HRegionServer
4986 DataNode
5125 TaskTracker


查看url：
http://192.168.40.240:60010/

参考
http://www.cnblogs.com/ventlam/archive/2011/01/22/HBaseCluster.html
http://hi.baidu.com/%C2%ED%B3%A4%D5%F72008/blog/item/0f1fdca390f527a2cbefd0de.html
http://wadefall.iteye.com/blog/1182803

问题：
http://javoft.net/2011/09/hbase-hmaster-%E6%97%A0%E6%B3%95%E5%90%AF%E5%8A%A8-call-to-failed-on-local-exception/
http://trophy.iteye.com/blog/1189119