hadoop2.7.3 HA高可用集群安装

hadoop2.7.3 HA高可用集群安装

1. 
   

   1. 环境准备

   2. Hadoop安装配置

   3. 
      

   4. 修改配置

   1. 全部节点安装jdk1.8，ssh，配置hosts文件，配置centos680和centos681之间的互相免登录，centos680到其余全部机器的免登录，关闭防火墙。

   2. 在centos682，centos683，centos684上安装Zookeeper3.4.9。

   3. 
      

   4. 复制hadoop-2.7.3.tar.gz到centos680的/opt目录下，之后全部的操做都在centos680上进行。

   5. 解压hadoop-2.7.3.tar.gz到/opt/bigdata文件夹下：tar –zxvf hadoop-2.7.3.tar.gz –C /opt/bigdata/。

   6. 
      

   1. 
      

   1. 
      

   1. ResourceManager借助zk实现热备，当某个节点失效以后另外一个节点可以被通知，并接管任务。

   1. 
      

   2. NN必须实现共享元数据才能保证无缝切换；共享元数据能够采用Linux提供的NFS服务，也能够使用Hadoop提供的JournalNode；JournalNode采用多数派理论，保证半数以上的节点写成功即认为成功。

   3. NN要实现自动切换（不须要手工切换），就必须实现实时监控每一个节点的状态；这里采用DFSFailoverController进行监控，若是某个NN失败，其余NN可以经过ZK被通知到，并接替NN的任务。

   1. HDFS高可用性（NameNode）

   3. RM高可用性（ResourceManager）

   5. Hadoop2.7.3 HA搭建步骤

2. 
   

   1. 
      

   2. hadoop-env.sh

   3. 
      

   4. 
      

3. 修改JAVA_HOME

4. 
   

   2. 
      

   3. core-site.xml

   4. 
      

   5. 
      

6. 
   

   3. 
      

   4. hdfs-site.xml

   5. 
      

   6. 
      

7. [html] view plain copy

   <property>
          <name>dfs.nameservices</name>
          <value>ns</value>
      </property>
   
      <property>
          <name>dfs.ha.namenodes.ns</name>
          <value>nn1,nn2</value>
      </property>
   
      <property>
          <name>dfs.namenode.rpc-address.ns.nn1</name>
          <value>centos680:9000</value>
      </property>
   
      <property>
          <name>dfs.namenode.http-address.ns.nn1</name>
          <value>centos680:50070</value>
      </property>
   
      <property>
          <name>dfs.namenode.rpc-address.ns.nn2</name>
          <value>centos681:9000</value>
      </property>
   
      <property>
          <name>dfs.namenode.http-address.ns.nn2</name>
          <value>centos681:50070</value>
      </property>
   
      <property>
          <name>dfs.namenode.shared.edits.dir</name>
          <value>qjournal://zk1:8485;zk2:8485;zk3:8485/ns</value>
      </property>
   
      <property>
          <name>dfs.journalnode.edits.dir</name>
          <value>/opt/big/hadoop-2.7.3/journaldata</value>
      </property>
   
      <property>
          <name>dfs.ha.automatic-failover.enabled</name>
          <value>true</value>
      </property>
   
      <property>
          <name>dfs.client.failover.proxy.provider.ns</name>
          <value>
                  org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider
          </value>
      </property>
      <property>
          <name>dfs.ha.fencing.methods</name>
          <value>
              sshfence
              shell(/bin/true)
          </value>
      </property>
      <property>
          <name>dfs.ha.fencing.ssh.private-key-files</name>
          <value>/root/.ssh/id_rsa</value>
      </property>
      <property>
          <name>dfs.ha.fencing.ssh.connect-timeout</name>
          <value>30000</value>
      </property>

8. 
   

   4. 
      

   5. mapred-side.xml
      

   6. 
      

   7. 
      

9. [html] view plain copy

   
   
   

   <span style="white-space:pre">  </span><property>
                   <name>mapreduce.framework.name</name>
                   <value>yarn</value>
           </property>

10. 
    

    5. 
       

    6. yarn-site.xml
       

    7. [html] view plain copy

       
       
       

    8. <configuration>  

    9.                                           <!--开启RM高可靠-->  

    10.   

    11.                                           <property>  

    12.   

    13.                                              <name>yarn.resourcemanager.ha.enabled</name>  

    14.   

    15.                                              <value>true</value>  

    16.   

    17.                                           </property>  

    18.   

    19.                                           <!--指定RM的cluster id -->  

    20.   

    21.                                           <property>  

    22.   

    23.                                              <name>yarn.resourcemanager.cluster-id</name>  

    24.   

    25.                                              <value>yrc</value>  

    26.   

    27.                                           </property>  

    28.   

    29.                                           <!--指定RM的名字-->  

    30.   

    31.                                           <property>  

    32.   

    33.                                              <name>yarn.resourcemanager.ha.rm-ids</name>  

    34.   

    35.                                              <value>rm1,rm2</value>  

    36.   

    37.                                           </property>  

    38.   

    39.                                           <!--分别指定RM的地址-->  

    40.   

    41.                                           <property>  

    42.   

    43.                                              <name>yarn.resourcemanager.hostname.rm1</name>  

    44.   

    45.                                              <value>h0</value>  

    46.   

    47.                                           </property>  

    48.   

    49.                                           <property>  

    50.   

    51.                                              <name>yarn.resourcemanager.hostname.rm2</name>  

    52.   

    53.                                              <value>h1</value>  

    54.   

    55.                                           </property>  

    56.   

    57.                                           <!--指定zk集群地址-->  

    58.   

    59.                                           <property>  

    60.   

    61.                                              <name>yarn.resourcemanager.zk-address</name>  

    62.   

    63.                                              <value>h2:2181,h3:2181,h4:2181</value>  

    64.   

    65.                                           </property>  

    66.   

    67.                                           <property>  

    68.   

    69.                                              <name>yarn.nodemanager.aux-services</name>  

    70.   

    71.                                              <value>mapreduce_shuffle</value>  

    72.   

    73.                                           </property>  

    74.   

    75.                             </configuration>  

    76. 
        

11. 
    

    6. 
       

    7. slaves
       

    8. 
       

    9. 
       

13. 
    

    2. 
       

    3. 分发到其余节点（在centos680上操做）
       

    4. 
       

14.   scp -r /opt/bigdata/hadoop-2.7.3/ h1:/opt/bigdata/
      scp -r /opt/bigdata/hadoop-2.7.3/ h2:/opt/bigdata/
      scp -r /opt/bigdata/hadoop-2.7.3/ h3:/opt/bigdata/
      scp -r /opt/bigdata/hadoop-2.7.3/ h4:/opt/bigdata/

15. 
    

    3. 
       

    4. 到h2,h3,h4启动全部的Zookeeper。

    5. 到h2,h3,h4启动JournalNode（hadoop-daemon.sh start journalnode）

    6. 到centos680上执行格式化namenode（hdfs namenode -format）,并将格式化后的元数据内容复制到另一个namenode节点中（h1）:scp -r tmp/ h1:/opt/bigdata/hadoop-2.7.3/(tmp是在core-site中配置的namenode元数据应该所处的位置)

    7. 格式化zk：bin/hdfs zkfc –formatZK。该操做在Zookeeper中建立了数据节点：
       

    8. 初始化和启动

    9. 
       

17. 
    

    5. 
       

    6. 启动dfs和yarn
       

    7. 
       

    8. 
       

18. sbin/start-dfs.sh

    sbin/start-yarn.sh

19. 
    

    4. 
       

    5. 验证
       

    6. 
       

20.   经过杀死active的namenode进程或者关闭namenode所在机器来验证切换状况。