1.4 集群搭建

解压安装，配置环境变量

   sudo chown -R admin:admin /opt/software/hadoop-2.7.2.tar.gz
   sudo chmod -R 755 /opt/software/hadoop-2.7.2.tar.gz
   tar -zxvf /opt/software/hadoop-2.7.2.tar.gz -C /opt/module
   cd /opt/module/hadoop-2.7.2
   sudo vim /etc/profile
   vim ~/.bashrc

# HADOOP_ENV
HADOOP_HOME=$BASE_DIR/hadoop-2.5.0-cdh5.3.6
PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin
export HADOOP_HOME PATH

   source /etc/profile
   source ~/.bashrc

修改配置

   vim etc/hadoop/core-site.xml

<!-- 定义hdfs的namenode节点 -->
   <property>
           <name>fs.defaultFS</name>
           <value>hdfs://hadoop102:9000</value>
   </property>
 <!-- hdfs 的是数据存放目录 -->
   <property>
           <name>hadoop.tmp.dir</name>
           <value>/opt/module/hadoop-2.7.2/data/tmp</value>
   </property>

       vim etc/hadoop/hdfs-site.xml

  <!-- 定义hdfs重要性略次于namenode节点的sn节点 -->
 <property>
           <name>dfs.namenode.secondary.http-address</name>
           <value>hadoop103:50090</value>
 </property>   
<!-- hdfs存放数据的副本数 -->
 <property>
           <name>dfs.replication</name>
           <value>3</value>
   </property>

       vim etc/hadoop/yarn-site.xml

<!-- RM 资源管理节点 -->
 <property>
         <name>yarn.resourcemanager.hostname</name>
         <value>hadoop104</value>
 </property> 
 
<!-- yarn 资源管理器洗牌策略 -->
 <property>
         <name>yarn.nodemanager.aux-services</name>
         <value>mapreduce_shuffle</value>
 </property>

       vim etc/hadoop/mapred-site.xml

 <!-- 声明 MR 运行在yarn上-->   
 <property>
        <name>mapreduce.framework.name</name>
        <value>yarn</value>
 </property>

       vim /opt/module/hadoop-2.7.2/etc/hadoop/slaves

hadoop102
hadoop103
hadoop104

集群分发

方案1:
   scp -r /opt/module/hadoop-2.7.2 hadoop103:/opt/module
   scp -r /opt/module/hadoop-2.7.2 hadoop104:/opt/module
方案2:
   sudo vim /usr/local/bin/xsync

#!/bin/bash

if (( $# == 0 ));then
  echo "no args"       
  exit 0
fi

USER=`whoami`
DIR=`cd -P $(dirname $1);pwd`
TARGET=`basename $1`
MYSELF=`hostname`

for (( host = 102 ; host <= 104 ; host++ ))
do
 if  [ $MYSELF != hadoop$host ];then
      echo "-------- Send $DIR/$TARGET to $USER@hadoop$host ---------"
      rsync -rvl $DIR/$TARGET $USER@hadoop$host:$DIR
 fi
done

exit 0

      sudo chown admin:admin /usr/local/bin/xsync
      sudo chmod 755 /usr/local/bin/xsync
      xsync /opt/module/hadoop-2.7.2

格式化

 [admin@hadoop102 hadoop-2.7.2]$ bin/hdfs namenode -format

启动服务

方案1：逐个节点启动
  sbin/hadoop-daemon.sh start|stop namenode|datanode|secondarynamenode
  sbin/yarn-daemon.sh start|stop resourcemanager|nodemanager
方案2：逐模块启动(Nn节点启动hdfs,Rm节点启动yarn)
  [admin@hadoop102 hadoop-2.7.2]$ sbin/start-dfs.sh
  [admin@hadoop102 hadoop-2.7.2]$ sbin/stop-dfs.sh
  [admin@hadoop104 hadoop-2.7.2]$ sbin/start-yarn.sh
  [admin@hadoop104 hadoop-2.7.2]$ sbin/stop-yarn.sh
方案3：群起
  [admin@hadoop102 hadoop-2.7.2]$ sbin/start-all.sh
  [admin@hadoop102 hadoop-2.7.2]$ sbin/stop-all.sh

查看运行状态

【jps 查看进程】
  [admin@hadoop102 hadoop-2.7.2]$ jps
【浏览器查看端口服务】
http://hadoop102:50070

hdfs.jpeg http://hadoop104:8088
yarn.jpeg