VirtualBox+Centos6.4搭建Hadoop1.1.2分布式环境
伪分布式环境搭建http://mvplee.iyunv.com/blog/2212629
复制之前在VirutalBox搭建的伪分布式环境，修改主机名，分别为hadoop1、hadoop2、hadoop3
修改每台主机的主机名，分别为hadoop1、hadoop2、hadoop3
/etc/sysconfig/network
/etc/hosts
reboot
修改hadoop1主机的配置文件
core-site.xml
mapred-site.xml
配置每台主机的ssh
# ssh-keygen -t rsa
# cd /root/.ssh
# cat id_rsa.pub >>authorized_keys
在hadoop2、hadoop3上复制自己的公钥到hadoop1上，hadoop1、hadoop2都可以免密码登录hadoop1
# ssh-copy-id id hadoop0
# ssh-copy-id -i hadoop1
# more authorized_keys
复制hadoop1中的authoried_keys到hadoop2、hadoop3上，确保三台主机上的authorized_keys内容一样都可以互相面密码登录
# scp /root/.ssh/authorized_keys hadoop2:/root/.ssh/
# scp /root/.ssh/authorized_keys hadoop3:/root/.ssh/
删除hadoop1上hadoop包中的logs、tmp目录
# rm -rf logs
# rm -rf tmp
在hadoop1上复制JDK目录和hadoop目录到hadoop1、hadoop2上
# scp -r /usr/local/jdk1.6.0_24 hadoop2:/usr/local/jdk1.6.0_24
# scp -r /usr/local/jdk1.6.0_24 hadoop3:/usr/local/jdk1.6.0_24
# scp -r /usr/local/hadoop-1.1.2 hadoop2:/usr/local/hadoop-1.1.2
# scp -r /usr/local/hadoop-1.1.2 hadoop3:/usr/local/hadoop-1.1.2
在hadoop1上复制profile到hadoop2、hadoop3上，重新加载资源文件
# scp /etc/profile hadoop2:/etc/
# scp /etc/profile hadoop3:/etc/
# source /etc/profile
# source /etc/profile
配置hadoop主从节点，在hadoop1主机上的slaves文件中加入hadoop2、hadoop3为从节点，masters上存放的是SecondarNameNode
# more hadoop-1.1.2/conf/slaves
#localhost
hadoop2
hadoop3
在hadoop1上格式namenode，启动hadoop
# hadoop namenode -format
# jps
5740 JobTracker
5659 SecondaryNameNode
5492 NameNode
5839 Jps
# jps
3473 TaskTracker
3568 Jps
3365 DataNode
# jps
2233 TaskTracker
2310 Jps
2142 DataNode
使用浏览器登录http://192.168.56.21:50070/
分布式搭建完成
页:
[1]