Hadoop学习-4-Hadoop部署（HA）

HA环境搭建

zookeeper集群搭建

• zookeeper由2n+1台服务器组成

• 集群搭建

  • java环境，需要安装jdk

  • 修稿配置文件

    # zoo.cfg
    dataDir=/opt/data/zookeeper_data
    
    server.1=node-01:2888:3888
    server.2=node-03:2888:3888
    server.3=node-03:2888:3888
    

  • 创建myid文件, echo 1 > /opt/data/zookeeper_data/myid

  • 分发到其他机器，并将zookeeper添加到环境变量中；修改myid为各自的id

  • 服务启动，zkServer.sh start

  • 服务状态、角色查询zkServer.sh status

  • 服务关闭zkServer.sh stop

hadoop集群

• 各个节点之间互相免密

• namenode中需要有zkfc，各个节点需要zk和jn

• 配置文件修改

  • Hdfs-env.sh

  # hdfs-evn.sh
  export JAVA_HOME=/opt/software/jdk1.8.0_311
  
  export HDFS_NAMENODE_USER=root
  export HDFS_DATANODE_USER=root
  export HDFS_SECONDARYNAMENODE_USER=root
  export YARN_RESOURCEMANAGER_USER=root
  export YARN_NODEMANAGER_USER=root
  export HDFS_ZKFC_USER=root
  export HDFS_JOURNALNODE_USER=root
  

  • Core-site.xml

  
  <configuration>
    	
      <property>
          <name>fs.defaultFSname>
          <value>hdfs://hadoop-clustervalue>
        	<description>集群名称，需要和hdfs-site.xml文件中保持一致description>
      property>
      <property>
          <name>hadoop.tmp.dirname>
          <value>/opt/data/hadoopvalue>
        	<description>临时文件夹，元数据默认文件夹description>
      property>
      <property>
          <name>hadoop.http.staticuser.username>
          <value>rootvalue>
        	<description>http请求用户description>
      property>
      <property>
          <name>hadoop.proxyuser.root.hostsname>
          <value>*value>
      property>
      <property>
          <name>hadoop.proxyuser.root.groupsname>
          <value>*value>
      property>
    	
      <property>
          <name>ha.zookeeper.quorumname>
          <value>node-01:2181,node-02:2181,node-03:2181value>
        	<description>zookeeper集群地址description>
      property>
  configuration>
  

  • hdfs-site.xml

  
  <configuration>
      
      <property>
          <name>dfs.nameservicesname>
          <value>hadoop-clustervalue>
      property>
      
      <property>
          <name>dfs.ha.namenodes.hadoop-clustername>
          <value>namenode-1, namenode-2value>
      property>
      
      <property>
          <name>dfs.namenode.rpc-address.hadoop-cluster.namenode-1name>
          <value>node-01:8020value>
      property>
      
      <property>
          <name>dfs.namenode.http-address.hadoop-cluster.namenode-1name>
          <value>node-01:9870value>
      property>
      
      <property>
          <name>dfs.namenode.rpc-address.hadoop-cluster.namenode-2name>
          <value>node-02:8020value>
      property>
      
      <property>
          <name>dfs.namenode.http-address.hadoop-cluster.namenode-2name>
          <value>node-02:9870value>
      property>
  		
      <property>
          <name>dfs.namenode.shared.edits.dirname>
          <value>qjournal://node-01:8485;node-02:8485;node-03:8485/hadoop-clustervalue>
      property>
      
      <property>
          <name>dfs.journalnode.edits.dirname>
          <value>/opt/data/hadoop/journaldatavalue>
      property>
      
    	<property>
    		<name>dfs.ha.automatic-failover.enabledname>
        <value>truevalue>
    	property>
    	
  		<property>
    		<name>dfs.client.failover.proxy.provider.hadoop-clustername>
        <value>org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvidervalue>
    	property>
    	
    	<property>
    		<name>dfs.ha.fencing.methodsname>
    		<value>sshfencevalue>
    	property>
    	
    	<property>
    		<name>dfs.ha.fencing.ssh.private-key-filesname>
        <value>/root/.ssh/id_rsavalue>
    	property>
    	
    	<property>
    		<name>dfs.ha.fencing.ssh.connect-timeoutname>
        <value>30000value>
    	property>
    	
      <property>
          <name>dfs.replicationname>
          <value>3value>
      property>
    	
      <property>
          <name>dfs.webhdfs.enabledname>
          <value>truevalue>
      property>
    	
      <property>
          <name>dfs.permissions.enabledname>
          <value>falsevalue>
      property>
    	
    	<property>
          <name>dfs.datanode.data.dirname>
          <value>/opt/data/datanodevalue>
      property>
    	
      <property>
          <name>dfs.ha.nn.not-become-active-in-safemodename>
          <value>falsevalue>
      property>
  configuration>
  

  • yarn高可用，修改yarn-site.xml

  
    <property>
      <name>yarn.resourcemanager.ha.enabledname>
      <value>truevalue>
    property>
  	<property>
      <name>yarn.resourcemanager.ha.automatic-failover.embeddedname>
      <value>truevalue>
      <description>ha状态切换为自动切换description>
  	property>
    
    <property>
      <name>yarn.resourcemanager.cluster-idname>
      <value>yarn-clustervalue>
    property>
    
    <property>
      <name>yarn.resourcemanager.ha.rm-idsname>
      <value>yarn-ha-rm1,yarn-ha-rm2value>
    property>
    
    <property>
      <name>yarn.resourcemanager.hostname.yarn-ha-rm1name>
      <value>node-01value>
    property>
    
    <property>
      <name>yarn.resourcemanager.hostname.yarn-ha-rm2name>
      <value>node-02value>
    property>  
    
    <property>
      <name>yarn.resourcemanager.ha.idname>
      <value>yarn-ha-rm1value>
    property>
  	
  	<property>
  		<name>yarn.resourcemanager.webapp.address.yarn-ha-rm1name>
      <value>node-01:8088value>
  	property>
  	<property>
  		<name>yarn.resourcemanager.webapp.address.yarn-ha-rm2name>
      <value>node-02:8088value>
  	property>
  	<property>
  		<name>yarn.resourcemanager.scheduler.address.yarn-ha-rm1name>
      <value>node-01:8030value>
  	property>
  	<property>
  		<name>yarn.resourcemanager.scheduler.address.yarn-ha-rm2name>
      <value>node-02:8030value>
  	property>
    
    <property>
      <name>yarn.resourcemanager.zk-addressname>
      <value>node-01:2181,node-02:2181,node-03:2181value>
    property> 
    
    <property>
      <name>yarn.nodemanager.aux-servicesname>
      <value>mapreduce_shufflevalue>
    property>
  	
  	<property>
  		<name>yarn.application.classpathname>
    	<value>/opt/server/hadoop/etc/hadoop:/opt/server/hadoop/share/hadoop/common/lib/*:/opt/server/hadoop/share/hadoop/common/*:/opt/server/hadoop/share/hadoop/hdfs:/opt/server/hadoop/share/hadoop/hdfs/lib/*:/opt/server/hadoop/share/hadoop/hdfs/*:/opt/server/hadoop/share/hadoop/mapreduce/lib/*:/opt/server/hadoop/share/hadoop/mapreduce/*:/opt/server/hadoop/share/hadoop/yarn:/opt/server/hadoop/share/hadoop/yarn/lib/*:/opt/server/hadoop/share/hadoop/yarn/*value>
      
  	property>
  

  • maprd.xml

  		<property>
          <name>mapreduce.framework.namename>
          <value>yarnvalue>
      property>
  

• 启动journalnode

  hdfs --daemon start journalnode

• 集群初始化hdfs namenode -format

• 启动namenodehdfs namenode start --daemon

• Node-02上同步元数据hdfs namenode -boostrapStandby

  • 或者执行scp -r /opt/data/hadoop/dfs root@node-02:/opt/data/hadoop

• 格式化zkfc，在active服务器上面执行即可 hdfs zkfc -formatZK

• 启动节点服务

  start-dfs.sh
  start-yarn.sh
  

• 如果切换失败，yum install psmisc -y

服务器断电重启集群

• 启动zookeeper集群，zkServer.sh start

• 启动journalnode, hdfs --daemon start journalnode

• 启动hdfs, yarn

  sh start-dfs.sh; sh start-yarn.sh