Hadoop集群部署初试

安装步骤:

  1.规划机器

  2.修改主机名称,配置ssh免登,安装jdk

  3.修改配置文件,创建目录

  4.启动应用

1.规划机器(centos1作为master)

   规划三台机器,一种centos1作为master,其余两台机器作为slaves

   10.240.139.101 centos1

   10.240.140.20   centos2

   10.240.139.72   centos3

   centos1安装NameNode SecondNameNode ResourceManager

   centos2安装DataNode NodeManager

2.修改主机名称,配置ssh免登,安装jdk

   [root@centos1 bin]# vi /etc/sysconfig/network

   NETWORKING=yes

   HOSTNAME=centos1

   NTPSERVERARGS=iburst

  

   [root@centos1 bin]# vi /etc/hosts

  10.240.139.101 centos1

  10.240.140.20  centos2

  10.240.139.72  centos3

   配置SSH免登,成功之后ssh localhost, ssh centos2不用用户名密码即成功。

   

#登录到centos1
cd ~/.ssh 
rm ./id_rsa*       #删除之前的key    
ssh-keygen -t rsa  #生成新的key,一直回车即可
cat ./id_rsa.pub >> ./authorized_keys
scp authorized_keys  root@centos2:~/.ssh/authorized_keys_from_centos1
#登录到centos2
cat authorized_keys_from_centos1 >> ./authorized_keys

   

    关闭防火墙:

sudo service iptables stop   # 关闭防火墙服务
sudo chkconfig iptables off  # 禁止防火墙开机自启,就不用手动关闭了

   JDK安装就滤过了。

3.修改配置文件

   hadoop-env.sh:

        hadoop的环境变量配置文件,需要配置JAVA_HOME的变量

   yarn-env.sh:  

       yarn的环境配置文件,需要配置JAVA_HOME的变量

   core-site.xml:

        hadoop的全局默认参数配置

   hdfs-site.xml:

        hdfs的参数配置

   yarn-site.xml:

        yarn的参数配置

   mapred-site.xml:

        mapred的参数配置

   slaves:

        从节点配置

    

    core-site.xml文件

<configuration>
    <property>
        <name>hadoop.tmp.dir</name>
        <value>file:/usr/local/hadoop/tmp</value>
        <description>Abase for other temporary directories.</description>
    </property>
    <property>
        <name>fs.defaultFS</name>
        <value>hdfs://centos1:9000</value>
    </property>
</configuration>

  

   hdfs-site.xml 

<configuration>
    <property>
        <name>dfs.namenode.secondary.http-address</name>
        <value>centos1:50090</value>
    </property>
    <property>
        <name>dfs.replication</name>
        <value>1</value>
    </property>
    <property>
        <name>dfs.namenode.name.dir</name>
        <value>file:/usr/local/hadoop/tmp/dfs/name</value>
    </property>
    <property>
        <name>dfs.datanode.data.dir</name>
        <value>file:/usr/local/hadoop/tmp/dfs/data</value>
    </property>
</configuration>

    slaves

centos2
centos3

    

    yarn-site.xml

<configuration>
   <property>
       <name>yarn.resourcemanager.hostname</name>
       <value>centos1</value>
   </property>
   <property>
       <name>yarn.nodemanager.aux-services</name>
       <value>mapreduce_shuffle</value>
   </property>
</configuration>

    mapred-site.xml

<configuration>
   <property>
       <name>mapreduce.framework.name</name>
       <value>yarn</value>
   </property>
   <property>
       <name>mapreduce.jobhistory.address</name>
       <value>centos1:10020</value>
   </property>
   <property>
       <name>mapreduce.jobhistory.webapp.address</name>
       <value>centos1:19888</value>
   </property>
</configuration>

   创建目录:/usr/local/hadoop/tmp/dfs/name /usr/local/hadoop/tmp/dfs/data  

4.启动应用

   ./bin/hdfs namenode -format

   ./sbin/start-dfs.sh    centos1上可以看到NameNode SecondaryNameNode  centos2,3上看到datanode

   ./start-yarn.sh centos1上会看到ResourceManager centos2,3上会看到NodeManager进程

查看hdfs相关:http://10.240.139.101:50090/

查看yarn相关:http://10.240.139.72:8042/node/node

   

相关推荐