1.1 分布结构 (jdk安装就不讲了,三台机器上都必须安装) 链接:hadoop1.1.2.tar.gz下载 :http://pan.baidu.com/s/1hrgPRH6 密码:s5d5
主节点(1个,是hadoop01):NameNode、JobTracker、SecondaryNameNode
从节点(2个,是hadoop02、hadoop03):DataNode、TaskTracker
1.2 各节点重新产生ssh加密文件
1.3编辑各个节点的/etc/hosts,在该文件中含有所有节点的ip与hostname的映射信息
[root@hadoop01conf]# vi /etc/hosts
127.0.0.1 localhost localhost.localdomain localhost4localhost4.localdomain4
::1 localhost localhost.localdomainlocalhost6 localhost6.localdomain6
192.168.255.111hadoop01
192.168.255.112hadoop02
192.168.255.113hadoop03
1.4 两两节点之间的SSH免密码登陆
ssh-copy-id -i hadoop01 或者 scp /root/.ssh/authorized_keys hadoop01:/root/.ssh/
1.5把hadoop01的hadoop目录下的logs和tmp删除
[root@hadoop01conf]# cd /usr/share/app/hadoop-1.1.2
[root@hadoop01hadoop-1.1.2]# rm -rf logs/
[root@hadoop01 hadoop-1.1.2]# rm -rf tmp/
1.6 修改hadoop01的配置文件 hadoop-env.sh 、hdfs-site.xml、core-site.xml、mapred-site.xml 4个配置文件
1.6.1.hadoop-env.sh 添加Jdk根目录
[root@hadoop01conf]# vi hadoop-env.sh
exportJAVA_HOME=/usr/local/java/jdk1.7.0_79(大概第9行)
1.6.2.hdfs-site.xml
[root@hadoop01conf]# vi hdfs-site.xml
<configuration>
<property>
<name>dfs.replication</name>
<value>1</value>
</property>
<property>
<name>dfs.permissions</name>
<value>false</value>
</property>
</configuration>
1.6.3.core-site.xml
[root@hadoop01conf]# vi core-site.xml
<configuration>
<property>
<name>fs.default.name</name>
<value>hdfs://hadoop01:9000</value>
</property>
<property>
<name>hadoop.tmp.dir</name>
<value>/usr/share/app/hadoop-1.1.2/tmp</value>
</property>
</configuration>
1.6.4.mapred-site.xml
[root@hadoop01conf]# vi mapred-site.xml
<configuration>
<property>
<name>mapred.job.tracker</name>
<value>hadoop01:9001</value>
</property>
</configuration>
1.7 文件安全拷贝
1.7.1 把hadoop01中的hadoop文件夹复制到hadoop02和hadoop03节点
[root@hadoop01hadoop-1.1.2]# scp -r /usr/share/app/hadoop-1.1.2 hadoop02:/usr/share/app/
[root@hadoop01 hadoop-1.1.2]# scp -r /usr/share/app/hadoop-1.1.2hadoop03:/usr/share/app/
1.7.2把hadoop0的/etc/profile复制到hadoop1和hadoop2节点,在目标节点中执行source /etc/profile
[root@hadoop01conf]# scp -r /etc/profile hadoop01:/etc/
[root@hadoop01conf]# source /etc/profile
1.8 编辑hadoop01中的 slaves,改为从节点的hostname,分别是hadoop02和hadoop03
[root@hadoop01conf]# vi slaves
hadoop02
hadoop03
~
1.9格式化,在hadoop01节点执行hadoopnamenode -format
1.10启动,在hadoop01节点执行start-all.sh
****注意:对于配置文件core-site.xml和mapred-site.xml在所有节点中都是相同的内容。
查看节点
主节点:hadoop01(NameNode、JobTracker、SecondaryNameNode),从节点hadoop02、hadoop03(DataNode、TaskTracker)上分别执行jps
[root@hadoop01 conf]# jps
2836 Jps
2594SecondaryNameNode
2438 NameNode
2671 JobTracker
[root@hadoop02hadoop-1.1.2]# jps
1880 TaskTracker
1969 Jps
1812 DataNode
[root@hadoop03hadoop-1.1.2]# jps
1819 DataNode
1904 TaskTracker
2008 Jps
和预想的一样,大功告成!