CentOS7搭建Hadoop-2.7集群

---

首先机器需要修改好静态ip

1.创建目录

mkdir -p /root/data/hadoop
mkdir -p /root/data/hadoop/tmp
mkdir -p /root/data/hadoop/var
mkdir -p /root/data/hadoop/dfs
mkdir -p /root/data/hadoop/dfs/name
mkdir -p /root/data/hadoop/dfs/data

2.修改/opt/hadoop-2.7/etc/hadoop/core-site.xml

    hadoop.tmp.dir
    /root/data/hadoop/tmp



     fs.default.name
     hdfs://hserver-1:9000



     hadoop.proxyuser.root.hosts
     *



     hadoop.proxyuser.root.groups
     *

3.修改/opt/hadoop-2.7/etc/hadoop/hadoop-env.sh

export JAVA_HOME=/opt/jdk-1.8

4.修改/opt/hadoop-2.7/etc/hadoop/hdfs-site.xml

    dfs.name.dir
    /root/data/hadoop/dfs/name



    dfs.data.dir
    /root/data/hadoop/dfs/data



    dfs.tmp.dir
    /root/data/hadoop/tmp



    dfs.replication
    2



    dfs.permissions
    false

5.修改mapred-site.xml文件

cp mapred.site.xml.template mapred.site.xml

    mapred.job.tracker
    hserver-1:49001



    mapred.local.dir
    /root/data/hadoop/var



    mapreduce.framework.name
    yarn

6.修改yarn-site.xml

    yarn.resourcemanager.hostname
    hserver-1



    yarn.nodemanager.aux-services
    mapreduce_shuffle

7.修改slaves文件

hserver-2
hserver-3

8.安装另外两台机器

需要从修改静态ip开始

在namenode格式化之前要实现三台机器的ssh免密登录