Hadoop HA高可用集群搭建

集群规划：

         主机名              IP                                  安装的软件                                          运行的进程

wm01       192.168.1.201          jdk、hadoop            NameNode、DFSZKFailoverController(zkfc)

wm02       192.168.1.202          jdk、hadoop            NameNode、DFSZKFailoverController(zkfc)

wm03       192.168.1.203          jdk、hadoop            ResourceManager

wm04       192.168.1.204          jdk、hadoop            ResourceManager

wm05       192.168.1.205          jdk、hadoop、zookeeper       DataNode、NodeManager、JournalNode、QuorumPeerMain

wmd06     192.168.1.206          jdk、hadoop、zookeeper       DataNode、NodeManager、JournalNode、QuorumPeerMain

wm07       192.168.1.207          jdk、hadoop、zookeeper       DataNode、NodeManager、JournalNode、QuorumPeerMain

 

HA集群搭建

1、  centos网络设置

http://blog.csdn.net/w0823m/article/details/55102138

2、  centos普通用户拥有sudo执行权限

http://blog.csdn.net/w0823m/article/details/55102332

3、  桌面版centos设置为server启动

http://blog.csdn.net/w0823m/article/details/55102475

4、  Centos修改主机名

http://blog.csdn.net/w0823m/article/details/55102813

5、  修改主机名和IP的映射关系

vim /etc/hosts

192.168.1.101           wm01

6、  关闭防火墙

           #查看防火墙状态

           service iptables status

           #关闭防火墙

           service iptables stop

           #查看防火墙开机启动状态

           chkconfig iptables --list

           #关闭防火墙开机启动

           chkconfig iptables off

7、  上传安装文件到服务器

8、  安装jdk

http://blog.csdn.net/w0823m/article/details/52599911

 

 

说明：

         1.在hadoop2.0中通常由两个NameNode组成，一个处于active状态，另一个处于standby状态。ActiveNameNode对外提供服务，而Standby NameNode则不对外提供服务，仅同步active namenode的状态，以便能够在它失败时快速进行切换。

         hadoop2.0官方提供了两种HDFSHA的解决方案，一种是NFS，另一种是QJM。这里我们使用简单的QJM。在该方案中，主备NameNode之间通过一组JournalNode同步元数据信息，一条数据只要成功写入多数JournalNode即认为写入成功。通常配置奇数个JournalNode

         这里还配置了一个zookeeper集群，用于ZKFC（DFSZKFailoverController）故障转移，当ActiveNameNode挂掉了，会自动切换Standby NameNode为standby状态

         2.hadoop-2.2.0中依然存在一个问题，就是ResourceManager只有一个，存在单点故障，hadoop-2.4.1解决了这个问题，有两个ResourceManager，一个是Active，一个是Standby，状态由zookeeper进行协调

安装步骤：

         1.安装配置zooekeeper集群（在wm05上）

                   1.1解压

                            tar-zxvf zookeeper-3.4.5.tar.gz -C /wm/

                   1.2修改配置

                            cd/wm/zookeeper-3.4.5/conf/

                            cpzoo_sample.cfg zoo.cfg

                            vimzoo.cfg

                            修改：dataDir=/wm/zookeeper-3.4.5/tmp

                            在最后添加：

                            server.1=wm05:2888:3888

                            server.2=wm06:2888:3888

                            server.3=wm07:2888:3888

                            保存退出

                            然后创建一个tmp文件夹

                            mkdir/wm/zookeeper-3.4.5/tmp

                            再创建一个空文件

                            touch/wm/zookeeper-3.4.5/tmp/myid

                            最后向该文件写入ID

                            echo1 > /wm/zookeeper-3.4.5/tmp/myid

                   1.3将配置好的zookeeper拷贝到其他节点(首先分别在wm06、wm07根目录下创建一个wm目录：mkdir /wm)

                            scp-r /wm/zookeeper-3.4.5/ wm06:/wm/

                            scp-r /wm/zookeeper-3.4.5/ wm07:/wm/

                           

                            注意：修改wm06、wm07对应/wm/zookeeper-3.4.5/tmp/myid内容

                            wm06：

                                     echo2 > /wm/zookeeper-3.4.5/tmp/myid

                            wm07：

                                     echo3 > /wm/zookeeper-3.4.5/tmp/myid

        

         2.安装配置hadoop集群（在wm01上操作）

                   2.1解压

                            tar-zxvf hadoop-2.4.1.tar.gz -C /wm/

                   2.2配置HDFS（hadoop2.0所有的配置文件都在$HADOOP_HOME/etc/hadoop目录下）

                            #将hadoop添加到环境变量中

                            vim/etc/profile

                            exportJAVA_HOME=/usr/java/jdk1.7.0_55

                            exportHADOOP_HOME=/wm/hadoop-2.4.1

                            exportPATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin

                           

                            #hadoop2.0的配置文件全部在$HADOOP_HOME/etc/hadoop下

                            cd/home/hadoop/app/hadoop-2.4.1/etc/hadoop

                           

                            2.2.1修改hadoo-env.sh

                                     exportJAVA_HOME=/home/hadoop/app/jdk1.7.0_55

                                    

                            2.2.2修改core-site.xml

                                     <configuration>

                                               <!--指定hdfs的nameservice为ns1 -->

                                               <property>

                                                        <name>fs.defaultFS</name>

                                                        <value>hdfs://ns1/</value>

                                               </property>

                                               <!--指定hadoop临时目录 -->

                                               <property>

                                                        <name>hadoop.tmp.dir</name>

                                                        <value>/home/hadoop/app/hadoop-2.4.1/tmp</value>

                                               </property>

                                              

                                               <!-- 指定zookeeper地址 -->

                                               <property>

                                                        <name>ha.zookeeper.quorum</name>

                                                        <value>wm05:2181,wm06:2181,wm07:2181</value>

                                               </property>

                                     </configuration>

                                    

                            2.2.3修改hdfs-site.xml

                                     <configuration>

                                               <!--指定hdfs的nameservice为ns1，需要和core-site.xml中的保持一致 -->

                                               <property>

                                                        <name>dfs.nameservices</name>

                                                        <value>ns1</value>

                                               </property>

                                               <!--ns1下面有两个NameNode，分别是nn1，nn2 -->

                                               <property>

                                                        <name>dfs.ha.namenodes.ns1</name>

                                                        <value>nn1,nn2</value>

                                               </property>

                                               <!--nn1的RPC通信地址 -->

                                               <property>

                                                        <name>dfs.namenode.rpc-address.ns1.nn1</name>

                                                        <value>wm01:9000</value>

                                               </property>

                                               <!--nn1的http通信地址 -->

                                               <property>

                                                        <name>dfs.namenode.http-address.ns1.nn1</name>

                                                        <value>wm01:50070</value>

                                               </property>

                                               <!--nn2的RPC通信地址 -->

                                               <property>

                                                        <name>dfs.namenode.rpc-address.ns1.nn2</name>

                                                        <value>wm02:9000</value>

                                               </property>

                                               <!--nn2的http通信地址 -->

                                               <property>

                                                        <name>dfs.namenode.http-address.ns1.nn2</name>

                                                        <value>wm02:50070</value>

                                               </property>

                                               <!--指定NameNode的元数据在JournalNode上的存放位置 -->

                                               <property>

                                                        <name>dfs.namenode.shared.edits.dir</name>

                                                        <value>qjournal://wm05:8485;wm06:8485;wm07:8485/ns1</value>

                                               </property>

                                               <!--指定JournalNode在本地磁盘存放数据的位置 -->

                                               <property>

                                                        <name>dfs.journalnode.edits.dir</name>

                                                        <value>/home/hadoop/app/hadoop-2.4.1/journaldata</value>

                                               </property>

                                               <!--开启NameNode失败自动切换 -->

                                               <property>

                                                        <name>dfs.ha.automatic-failover.enabled</name>

                                                        <value>true</value>

                                               </property>

                                               <!--配置失败自动切换实现方式 -->

                                               <property>

                                                        <name>dfs.client.failover.proxy.provider.ns1</name>

                                                        <value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider</value>

                                               </property>

                                               <!--配置隔离机制方法，多个机制用换行分割，即每个机制暂用一行-->

                                               <property>

                                                        <name>dfs.ha.fencing.methods</name>

                                                        <value>

                                                                 sshfence

                                                                 shell(/bin/true)

                                                        </value>

                                               </property>

                                               <!--使用sshfence隔离机制时需要ssh免登陆 -->

                                               <property>

                                                        <name>dfs.ha.fencing.ssh.private-key-files</name>

                                                        <value>/home/hadoop/.ssh/id_rsa</value>

                                               </property>

                                               <!--配置sshfence隔离机制超时时间 -->

                                               <property>

                                                        <name>dfs.ha.fencing.ssh.connect-timeout</name>

                                                        <value>30000</value>

                                               </property>

                                     </configuration>

                           

                            2.2.4修改mapred-site.xml

                                     <configuration>

                                               <!--指定mr框架为yarn方式 -->

                                               <property>

                                                        <name>mapreduce.framework.name</name>

                                                        <value>yarn</value>

                                               </property>

                                     </configuration>     

                           

                            2.2.5修改yarn-site.xml

                                     <configuration>

                                                        <!--开启RM高可用 -->

                                                        <property>

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

                                                           <value>true</value>

                                                        </property>

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

                                                        <property>

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

                                                           <value>yrc</value>

                                                        </property>

                                                        <!--指定RM的名字 -->

                                                        <property>

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

                                                           <value>rm1,rm2</value>

                                                        </property>

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

                                                        <property>

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

                                                           <value>wm03</value>

                                                        </property>

                                                        <property>

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

                                                           <value>wm04</value>

                                                        </property>

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

                                                        <property>

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

                                                          <value>wm05:2181,wm06:2181,wm07:2181</value>

                                                        </property>

                                                        <property>

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

                                                           <value>mapreduce_shuffle</value>

                                                        </property>

                                     </configuration>

                           

                                    

                            2.2.6修改slaves(slaves是指定子节点的位置，因为要在wm01上启动HDFS、在wm03启动yarn，所以wm01上的slaves文件指定的是datanode的位置，wm03上的slaves文件指定的是nodemanager的位置)

                                     wm05

                                     wm06

                                     wm07

 

                            2.2.7配置免密码登陆

                                     #首先要配置wm01到wm02、wm03、wm04、wm05、wm06、wm07的免密码登陆

                                     #在wm01上生产一对钥匙

                                     ssh-keygen-t rsa

                                     #将公钥拷贝到其他节点，包括自己

                                     ssh-coyp-idwm01

                                     ssh-coyp-idwm02

                                     ssh-coyp-idwm03

                                     ssh-coyp-idwm04

                                     ssh-coyp-idwm05

                                     ssh-coyp-idwm06

                                     ssh-coyp-idwm07

                                     #配置wm03到wm04、wm05、wm06、wm07的免密码登陆

                                     #在wm03上生产一对钥匙

                                     ssh-keygen-t rsa

                                     #将公钥拷贝到其他节点

                                     ssh-coyp-idwm04

                                     ssh-coyp-idwm05

                                     ssh-coyp-idwm06

                                     ssh-coyp-idwm07

                                     #注意：两个namenode之间要配置ssh免密码登陆，别忘了配置wm02到wm01的免登陆

                                     在wm02上生产一对钥匙

                                     ssh-keygen-t rsa

                                     ssh-coyp-id-i wm01                           

                  

                   2.4将配置好的hadoop拷贝到其他节点

                            scp-r /wm/ wm02:/

                            scp-r /wm/ wm03:/

                            scp-r /wm/hadoop-2.4.1/ hadoop@wm04:/wm/

                            scp-r /wm/hadoop-2.4.1/ hadoop@wm05:/wm/

                            scp-r /wm/hadoop-2.4.1/ hadoop@wm06:/wm/

                            scp-r /wm/hadoop-2.4.1/ hadoop@wm07:/wm/

                   ###注意：严格按照下面的步骤

                   2.5启动zookeeper集群（分别在wm05、wm06、tcast07上启动zk）

                            cd/wm/zookeeper-3.4.5/bin/

                            ./zkServer.shstart

                            #查看状态：一个leader，两个follower

                            ./zkServer.shstatus

                           

                   2.6启动journalnode（分别在在wm05、wm06、tcast07上执行）

                            cd/wm/hadoop-2.4.1

                            sbin/hadoop-daemon.shstart journalnode

                            #运行jps命令检验，wm05、wm06、wm07上多了JournalNode进程

                  

                   2.7格式化HDFS

                            #在wm01上执行命令:

                            hdfsnamenode -format

                            #格式化后会在根据core-site.xml中的hadoop.tmp.dir配置生成个文件，这里我配置的是/wm/hadoop-2.4.1/tmp，然后将/wm/hadoop-2.4.1/tmp拷贝到wm02的/wm/hadoop-2.4.1/下。

                            scp-r tmp/ wm02:/home/hadoop/app/hadoop-2.4.1/

                            ##也可以这样，建议hdfsnamenode -bootstrapStandby

                  

                   2.8格式化ZKFC(在wm01上执行即可)

                            hdfszkfc -formatZK

                  

                   2.9启动HDFS(在wm01上执行)

                            sbin/start-dfs.sh

 

                   2.10启动YARN(#####注意#####：是在wm03上执行start-yarn.sh，把namenode和resourcemanager分开是因为性能问题，因为他们都要占用大量资源，所以把他们分开了，他们分开了就要分别在不同的机器上启动)

                            sbin/start-yarn.sh

 

                  

         到此，hadoop-2.4.1配置完毕，可以统计浏览器访问:

                   http://192.168.1.201:50070

                   NameNode'wm01:9000' (active)

                   http://192.168.1.202:50070

                   NameNode'wm02:9000' (standby)

        

         验证HDFS HA

                   首先向hdfs上传一个文件

                   hadoopfs -put /etc/profile /profile

                   hadoopfs -ls /

                   然后再kill掉active的NameNode

                   kill-9 <pid of NN>

                   通过浏览器访问：http://192.168.1.202:50070

                   NameNode'wm02:9000' (active)

                   这个时候wm02上的NameNode变成了active

                   在执行命令：

                   hadoopfs -ls /

                   -rw-r--r--   3 root supergroup       1926 2014-02-06 15:36 /profile

                   刚才上传的文件依然存在！！！

                   手动启动那个挂掉的NameNode

                   sbin/hadoop-daemon.shstart namenode

                   通过浏览器访问：http://192.168.1.201:50070

                   NameNode'wm01:9000' (standby)

        

         验证YARN：

                   运行一下hadoop提供的demo中的WordCount程序：

                   hadoopjar share/hadoop/mapreduce/hadoop-mapreduce-examples-2.4.1.jar wordcount/profile /out

        

         OK，大功告成！！！