1.準(zhǔn)備zookeeper服務(wù)器

#node1,node2,node3
#安裝請參考http://suyanzhu.blog.51cto.com/8050189/1946580

創(chuàng)新互聯(lián)建站為您提適合企業(yè)的網(wǎng)站設(shè)計(jì)?讓您的網(wǎng)站在搜索引擎具有高度排名，讓您的網(wǎng)站具備超強(qiáng)的網(wǎng)絡(luò)競爭力！結(jié)合企業(yè)自身，進(jìn)行網(wǎng)站設(shè)計(jì)及把握，最后結(jié)合企業(yè)文化和具體宗旨等，才能創(chuàng)作出一份性化解決方案。從網(wǎng)站策劃到成都網(wǎng)站建設(shè)、成都網(wǎng)站設(shè)計(jì)， 我們的網(wǎng)頁設(shè)計(jì)師為您提供的解決方案。

2.準(zhǔn)備NameNode節(jié)點(diǎn)

#node1,node4

3.準(zhǔn)備JournalNode節(jié)點(diǎn)

#node2,node3,node4

4.準(zhǔn)備DataNode節(jié)點(diǎn)

#node2,node3,node4
#啟動DataNode節(jié)點(diǎn)命令hadoop-daemon.sh start datanode

5.修改hadoop的hdfs-site.xml配置文件

        
                dfs.nameservices
                yunshuocluster
        
        
                dfs.ha.namenodes.yunshuocluster
                nn1,nn2
        
        
                dfs.namenode.rpc-address.yunshuocluster.nn1
                node1:8020
        
        
                dfs.namenode.rpc-address.yunshuocluster.nn2
                node4:8020
        
        
                dfs.namenode.http-address.yunshuocluster.nn1
                node1:50070
        
        
                dfs.namenode.http-address.yunshuocluster.nn2
                node4:50070
        
        
                dfs.namenode.shared.edits.dir
                qjournal://node2:8485;node3:8485;node4:8485/yunshuocluste
r
        
        
                dfs.client.failover.proxy.provider.mycluster
                org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailo
verProxyProvider
        
        
                dfs.ha.fencing.methods
                sshfence
        
        
                dfs.ha.fencing.ssh.private-key-files
                /root/.ssh/id_dsa
        
        
                dfs.journalnode.edits.dir
                /opt/journalnode/
        
        
                dfs.ha.automatic-failover.enabled
                true
        

6.修改hadoop的core-site.xml配置文件

    
        fs.defaultFS
        hdfs://yunshuocluster
    
    
        hadoop.tmp.dir
        /opt/hadoop-2.5
    
    
        ha.zookeeper.quorum
        node1:2181,node2:2181,node3:2181
    

7.配置slaves配置文件

node2
node3
node4

8.啟動zookeeper(node1,node2,node3)

zkServer.sh start

9.啟動Journalnode(node2,node3,node4上分別執(zhí)行下面的命令)

#啟動命令 停止命令hadoop-daemon.sh stop journalnode
hadoop-daemon.sh start journalnode

10.檢查Journalnode,通過查看日志

cd /home/hadoop-2.5.1/logs
ls
tail -200 hadoop-root-journalnode-node2.log

11.格式化NameNode(兩臺中的一臺,這里格式化node4這臺NameNode節(jié)點(diǎn))

hdfs namenode -format

cd /opt/hadoop-2.5
#兩臺NameNode同步完成
scp -r /opt/hadoop-2.5/* root@node1:/opt/hadoop-2.5/

12.初始化zkfc

hdfs zkfc -formatZK

13.啟動服務(wù)

start-dfs.sh
#stop-dfs.sh表示停止服務(wù)