hadoop集群安装

环境：

Centos7.0，JDK1.8

设置主机名

hostname  hadoop1  hadoop2  hadoop3

hosts

/etc/hosts

192.168.31.17  hadoop1

192.168.31.210  hadoop2

192.168.31.65  hadoop3

免密钥登陆

ssh-keygen -t  rsa

ssh-copy-id -i  hadoop1

ssh-copy-id -i  hadoop2

ssh-copy-id -i  hadoop3

关闭防火墙

systemctl disable firewalld

yum remove firewalld -y

下载并解压hadoop,conf下修改

1.hadoop-env.sh

export JAVA_HOME=/usr/local/jdk/

2.core-site.xml

3.hdfs-site.xml

4.mapred-site.xml

5.masters    hadoop1

6.slaves    hadoop2  hadoop3

7.传到其它服务器

scp  -r  ./hadoop-1.1.2  hadoop2:/root

scp  -r  ./hadoop-1.1.2  hadoop3:/root

8.运行hadoop

hadoop的bin目录下运行：hadoop  namenode  -format

hadoop的bin目录下运行：./start-all.sh

简单例子测试

创建两个数据文件

echo  "hello  world"  >test1.txt

echo  "hello  hadoop"  >test2.txt

导入数据文件，运行hadoop example例子

bin/hadoop  dfs  -put  ../input    in

bin/hadoop  jar hadoop-examples-1.1.2.jar  wordcount  in  out

查看运行结果

bin/hadoop  dfs  -cat  ./outs/*

web地址

http://192.168.31.17:50030/jobtracker.jsp