Hadoop 完全分布式搭建

搭建环境

https://www.cnblogs.com/YuanWeiBlogger/p/11456623.html

修改主机名
-------------------
1./etc/hostname
　　s129
2./etc/hosts
　　127.0.0.1 localhost
　　192.168.248.129 s129
　　192.168.248.128 s128
　　192.168.248.127 s127
　　192.168.248.126 s126

完全分布式

1.克隆3台client(centos6.8)
右键centos-->管理->克隆-> ... -> 完整克隆
2.启动client
3.启用客户机共享文件夹。
4.修改hostname和ip地址文件
[/etc/hostname]
s127

[/etc/sysconfig/network-scripts/ifcfg-ethxxxx]
...
IPADDR=..

5.重启网络服务
$>sudo service network restart

6.修改/etc/resolv.conf文件
nameserver 192.168.231.2

7.重复以上3 ~ 6过程.

准备完全分布式主机的ssh
-------------------------
1.删除所有主机上的/home/yw/.ssh/*

2.在s129主机上生成密钥对
$>ssh-keygen -t rsa -P '' -f ~/.ssh/id_rsa

 

 

 

3.将s129的公钥文件id_rsa.pub远程复制到126 ~128主机上。
并放置/home/yw/.ssh/authorized_keys

$>scp id_rsa.pub yw@s129:/home/yw/.ssh/authorized_keys

$>scp id_rsa.pub yw@s128:/home/yw/.ssh/authorized_keys
$>scp id_rsa.pub yw@s127:/home/yw/.ssh/authorized_keys
$>scp id_rsa.pub yw@s126:/home/yw/.ssh/authorized_keys

4.配置完全分布式(${hadoop_home}/etc/hadoop/)

 

 

 修改完全分布式的xml文件

 

[core-site.xml]

fs.defaultFS
hdfs://s129/

[hdfs-site.xml]

dfs.replication
3



[mapred-site.xml]
不变

[yarn-site.xml]

 

 

 

yarn.resourcemanager.hostname
s129


yarn.nodemanager.aux-services
mapreduce_shuffle

[slaves]
s128
s127
s126

 

 

[hadoop-env.sh]
...
export JAVA_HOME=/opt/jdk

 

 

 

 

...

5.分发配置

$>cd /usr/local/hadoop/hadoop-2.7.3/etc

$>scp -r full-hadoop root@s128:/usr/local/hadoop/hadoop-2.7.3/etc/

$>scp -r full-hadoop root@s127:/usr/local/hadoop/hadoop-2.7.3/etc/

$>scp -r full-hadoop root@s126:/usr/local/hadoop/hadoop-2.7.3/etc/

 

6.删除符号连接 (把原来的伪分布式连接删掉)

$>cd /usr/local/hadoop/hadoop-2.7.3/etc
$>rm -rf hadoop

$>ssh s128 rm -rf /usr/local/hadoop/hadoop-2.7.3/etc/hadoop/

$>ssh s127 rm -rf /usr/local/hadoop/hadoop-2.7.3/etc/hadoop/

$>ssh s126 rm -rf /usr/local/hadoop/hadoop-2.7.3/etc/hadoop/

 

 

7.创建符号连接
$>cd /usr/local/hadoop/hadoop-2.7.3/etc
$>ln -s full-hadoop hadoop

$>ssh s126 ln -s /usr/local/hadoop/hadoop-2.7.3/etc/full-hadoop /usr/local/hadoop/hadoop-2.7.3/etc/hadoop

$>ssh s127 ln -s /usr/local/hadoop/hadoop-2.7.3/etc/full-hadoop /usr/local/hadoop/hadoop-2.7.3/etc/hadoop

$>ssh s128 ln -s /usr/local/hadoop/hadoop-2.7.3/etc/full-hadoop /usr/local/hadoop/hadoop-2.7.3/etc/hadoop

 

 

8.删除临时目录文件
$>cd /tmp
$>rm -rf hadoop-centos
$>ssh s128 rm -rf /tmp/hadoop-centos
$>ssh s127 rm -rf /tmp/hadoop-centos
$>ssh s126 rm -rf /tmp/hadoop-centos

 

 

9.删除hadoop日志

$>cd /usr/local/hadoop/hadoop-2.7.3/logs
$>rm -rf *
$>ssh s128 rm -rf /usr/local/hadoop/hadoop-2.7.3/logs/*
$>ssh s127 rm -rf /usr/local/hadoop/hadoop-2.7.3/logs/*
$>ssh s126 rm -rf /usr/local/hadoop/hadoop-2.7.3/logs/*

 

 

10.格式化文件系统
$>hadoop namenode -format

11.启动hadoop进程
$>start-all.sh

 

 

 

完全分布式搭建完成。