> Linux集群 > Hadoop >

Hadoop2.x下安装HBase

环境:CentOS6.5  Hadoop2.5.2  HBase1.0.0 

1.安装好 hadoop 集群,并启动 
[grid@hadoop4 ~]$ sh hadoop-2.5.2/sbin/start-dfs.sh 
[grid@hadoop4 ~]$ sh hadoop-2.5.2/sbin/start-yarn.sh 
查看 hadoop 版本: 
[grid@hadoop4 ~]$ hadoop-2.5.2/bin/hadoop version 
Hadoop 2.5.2 
Subversion Unknown -r Unknown 
Compiled by grid on 2014-12-31T01:40Z 
Compiled with protoc 2.5.0 
From source with checksum df7537a4faa4658983d397abf4514320 
This command was run using /home/grid/hadoop-2.5.2/share/hadoop/common/hadoop-common-2.5.2.jar 

2.查看 hbase 官方文档(http://hbase.apache.org/book.html#basic.prerequisites),找到与 hadoop 版本对应的 hbase 并下载 
[grid@hadoop4 ~]$ wget http://mirrors.cnnic.cn/apache/hbase/hbase-1.0.0/hbase-1.0.0-bin.tar.gz 

3.解压 
[grid@hadoop4 ~]$ tar -zxf hbase-1.0.0-bin.tar.gz 

4.进入 hbase 的 lib 目录,查看 hadoop jar 包的版本 
[grid@hadoop4 ~]$ cd hbase-1.0.0/lib/ 
[grid@hadoop4 lib]$ find -name 'hadoop*jar' 
./hadoop-common-2.5.1.jar 
./hadoop-mapreduce-client-common-2.5.1.jar 
./hadoop-annotations-2.5.1.jar 
./hadoop-yarn-server-common-2.5.1.jar 
./hadoop-hdfs-2.5.1.jar 
./hadoop-client-2.5.1.jar 
./hadoop-mapreduce-client-shuffle-2.5.1.jar 
./hadoop-yarn-common-2.5.1.jar 
./hadoop-yarn-server-nodemanager-2.5.1.jar 
./hadoop-yarn-client-2.5.1.jar 
./hadoop-mapreduce-client-core-2.5.1.jar 
./hadoop-auth-2.5.1.jar 
./hadoop-mapreduce-client-app-2.5.1.jar 
./hadoop-yarn-api-2.5.1.jar 
./hadoop-mapreduce-client-jobclient-2.5.1.jar 
发现与 hadoop 集群的版本号不一致,需要用 hadoop 目录下的 jar 替换 hbase/lib 目录下的 jar 文件。 
编写脚本来完成替换,如下所示: 
[grid@hadoop4 lib]$ pwd
/home/grid/hbase-1.0.0/lib
[grid@hadoop4 lib]$ vim f.sh
find -name "hadoop*jar" | sed 's/2.5.1/2.5.2/g' | sed 's/\.\///g' > f.log
rm ./hadoop*jar
cat ./f.log | while read Line
do
find /home/grid/hadoop-2.5.2 -name "$Line" | xargs -i cp {} ./
done
rm ./f.log
[grid@hadoop4 lib]$ chmod u+x f.sh
[grid@hadoop4 lib]$ ./f.sh
[grid@hadoop4 lib]$ find -name 'hadoop*jar'
./hadoop-yarn-api-2.5.2.jar
./hadoop-mapreduce-client-app-2.5.2.jar
./hadoop-common-2.5.2.jar
./hadoop-mapreduce-client-jobclient-2.5.2.jar
./hadoop-mapreduce-client-core-2.5.2.jar
./hadoop-yarn-server-nodemanager-2.5.2.jar
./hadoop-hdfs-2.5.2.jar
./hadoop-yarn-common-2.5.2.jar
./hadoop-mapreduce-client-shuffle-2.5.2.jar
./hadoop-auth-2.5.2.jar
./hadoop-mapreduce-client-common-2.5.2.jar
./hadoop-yarn-client-2.5.2.jar
./hadoop-annotations-2.5.2.jar
./hadoop-yarn-server-common-2.5.2.jar

OK,jar 包替换成功;hbase/lib 目录下还有个 slf4j-log4j12-XXX.jar,在机器有装hadoop时,由于classpath中会有hadoop中的这个jar包,会有冲突,直接删除掉

[grid@hadoop4 lib]$ rm `find -name 'slf4j-log4j12-*jar'`


5.修改配置文件

5.1. 
[grid@hadoop4 hbase-1.0.0]$ vi conf/hbase-env.sh 
export JAVA_HOME=/usr/java/jdk1.7.0_72 
export HBASE_CLASSPATH=/home/grid/hadoop-2.5.2/etc/hadoop 
export HBASE_MANAGES_ZK=true 
第一个参数指定了JDK路径;第二个参数指定了 hadoop 的配置文件路径;第三个参数设置使用 hbase 默认自带的 Zookeeper 

5.2. 
[grid@hadoop4 hbase-1.0.0]$ vim conf/hbase-site.xml 
<property>
<name>hbase.rootdir</name>
<value>hdfs://hadoop4:9000/hbase</value>
</property>
<property>
<name>hbase.cluster.distributed</name>
<value>true</value>
</property>
<property>
<name>hbase.tmp.dir</name>
<value>/home/grid/hbase-1.0.0/tmp</value>
</property>
<property>
<name>hbase.zookeeper.quorum</name>
<value>hadoop4,hadoop5,hadoop6</value>
</property>
<property>
<name>hbase.zookeeper.property.dataDir</name>
<value>/home/grid/hbase-1.0.0/zookeeper</value>
</property>

创建目录

[grid@hadoop4 hbase-1.0.0]$ mkdir tmp 
[grid@hadoop4 hbase-1.0.0]$ mkdir zookeeper 

5.3. 
[grid@hadoop4 hbase-1.0.0]$ vim conf/regionservers 
hadoop4 
hadoop5 
hadoop6 

6.设置环境变量 
[grid@hadoop4 ~]$ vi .bash_profile 
export HBASE_HOME=/home/grid/hbase-1.0.0 
export PATH=$PATH:$HBASE_HOME/bin 
[grid@hadoop4 ~]$ source .bash_profile 

7.分发 hbase 到其它机器,并在其上设置环境变量 
[grid@hadoop4 ~]$ scp -r hbase-1.0.0 grid@hadoop5:~ 
[grid@hadoop4 ~]$ scp -r hbase-1.0.0 grid@hadoop6:~ 

8.启动 hbase 
[grid@hadoop4 ~]$ sh start-hbase.sh 
[grid@hadoop4 ~]$ jps 
2388 ResourceManager 
3692 Jps 
2055 NameNode 
3375 HQuorumPeer 
2210 SecondaryNameNode 
3431 HMaster 
[grid@hadoop5 ~]$ jps 
2795 Jps 
2580 HQuorumPeer 
2656 HRegionServer 
2100 NodeManager 
1983 DataNode 
[grid@hadoop6 ~]$ jps 
2566 HQuorumPeer 
1984 DataNode 
2101 NodeManager 
2803 Jps 
2639 HRegionServer 
jps查看发现 Master 机上 HRegionServer 服务未启动,查看日志显示因16020端口被占用导致 HRegionServer 启动失败,查证发现占用16020端口的是 HMaster 进程,查看官方文档后解决:[grid@hadoop4 ~]$ sh local-regionservers.sh start 2 
官方文档截图: 
 

9.shell 
 

10.Web管理界面 
 



(责任编辑:IT)