官网手册：

http://hadoop.apache.org/docs/r1.0.4/cn/hdfs_design.html


主机信息

主机名                                 ip   

hadoop1                              10.0.70.242

hadoop2                              10.0.70.243

hadoop3                              10.0.70.230

hadoop4                              10.0.70.231

 

一．设置主机名映射

# vim /etc/hosts

127.0.0.1  localhost localhost.localdomain localhost4 localhost4.localdomain4

::1        localhost localhost.localdomain localhost6 localhost6.localdomain6

10.0.70.242     hadoop1

10.0.70.243     hadoop2

10.0.70.230     hadoop3

10.0.70.231     hadoop4

并拷贝到其他3台机器。

 

二．配置免密码登录

# cd /root/.ssh/

# ssh-keygen -t rsa

# cat id_rsa.pub >> authorized_keys

将其他3台机器的id_rsa.pub内容追加到authorized_keys文件中，并拷贝给其他3台机器。

#ssh hadoop1（hadoop2、hadoop3、hadoop4）进行验证。

 

三．安装jdk

# tar zxvf jdk-8u51-linux-x64.tar.gz -C/app/zpy/

配置环境变量

# vim /etc/profile

JAVA_HOME=/app/zpy/jdk1.8.0_51

JAVA_BIN=/app/zpy/jdk1.8.0_51/bin

PATH=$PATH:$JAVA_BIN

export JAVA_HOME JAVA_BIN PATH

# . /etc/profile

# java –version         进行验证

其他3台机器相同操作。

 

四．安装hadoop

# cd /app/zpy/3rd

# wget http://mirror.bit.edu.cn/apache/hadoop/common/hadoop-2.7.3/hadoop-2.7.3.tar.gz

# tar zxvf hadoop-2.7.3.tar.gz -C /app/zpy/

配置环境变量

# vim /etc/profile

HADOOP_HOME=/app/zpy/hadoop-2.7.3

HADOOP_BIN=/app/zpy/hadoop-2.7.3/bin

PATH=$PATH:$JAVA_BIN:$HADOOP_HOME:$HADOOP_BIN

export JAVA_HOME JAVA_BIN PATH HADOOP_HOMEHADOOP_BIN

# . /etc/profile

修改主要配置文件

# cd /app/zpy/hadoop-2.7.3/etc/hadoop

 

# vim hadoop-env.sh

添加

export JAVA_HOME=/app/zpy/jdk1.8.0_51

export HADOOP_HOME_WARN_SUPPRESS=1

# vim yarn-env.sh

更改

JAVA_HOME=/app/zpy/jdk1.8.0_51

 

# vim core-site.xml

内容如下

   

       fs.defaultFS

       hdfs://hadoop1:9000

   

 

   

       hadoop.tmp.dir

       /data/tmp

   

   

       io.file.buffer.size

       131702

   

 

# vim hdfs-site.xml

内容如下

   

       dfs.namenode.name.dir

       file:/data/hdfs/name/

   

 

   

       dfs.datanode.data.dir

       file:/data/hdfs/data

   

   

       dfs.replication

       2

   

 

   

       dfs.namenode.secondary.http-address

       hadoop1:9001

   

 

   

       dfs.webhdfs.enabled

       true

   

# cp mapred-site.xml.templatemapred-site.xml

 

# vim mapred-site.xml

内容如下

       mapreduce.framework.name

       yarn

   

# vim yarn-site.xml

内容如下

   

       yarn.nodemanager.aux-services

       mapreduce_shuffle

   

   

       yarn.nodemanager.auxservices.mapreduce.shuffle.class

       org.apache.hadoop.mapred.ShuffleHandler

   

   

       yarn.resourcemanager.address

       hadoop1:8032

   

   

       yarn.resourcemanager.scheduler.address

       hadoop1:8030

   

   

       yarn.resourcemanager.resource-tracker.address

       hadoop1:8031

   

   

       yarn.resourcemanager.admin.address

       hadoop1:8033

   

   

       yarn.resourcemanager.webapp.address

       hadoop1:8088

   

   

       yarn.scheduler.maximum-allocation-mb

       3000

   

 

 

# vim slaves

内容如下

hadoop1

hadoop2

hadoop3

hadoop4

 

创建必须的文件夹

 

# mkdir -p /data/tmp/

 #mkdir /data/hdfs

# mkdir /data/hdfs/data

# mkdir /data/hdfs/name

 

五．将配置好的文件拷贝到其他主机

# scp –r hadoop-2.7.3 hadoop2:/app/zpy

# scp –r hadoop-2.7.3 hadoop3:/app/zpy

# scp –r hadoop-2.7.3 hadoop4:/app/zpy

 

# scp /etc/profile hadoop2:/etc

# scp /etc/profile hadoop3:/etc

# scp /etc/profile hadoop4:/etc

 

刷新另外3台服务器的环境变量。

 

六．启动

首先格式化namenode节点，注意在master节点上！

# cd /app/zpy/hadoop-2.7.3/bin/

# hadoop namenode –format

 

然后启动集群

# start-dfs.sh

# start-yarn.sh

 

在master节点上运行

# jps

 

看到如下结果：

4083 Jps

30084 DataNode

30261 SecondaryNameNode

30550 NodeManager

29945 NameNode

30430 ResourceManager

 

切换到slave节点，运行

# jps

 

看到如下结果：

18561 NodeManager

4362 Jps

18443 DataNode

 

浏览器中打开http://10.0.70.242:50070/dfshealth.html#tab-overview可看见相应信息

 

七．运行测试用例

切换到hadoop根目录

执行如下命令解除ｈａｄｏｏｐ安全模式：

# ./bin/hadoop dfsadmin -safemode leave

 

再执行：

# ./bin/hadoop jarshare/hadoop/mapreduce/hadoop-mapreduce-examples-2.7.3.jar pi 1 1

 

显示如下结果：

Number of Maps  = 1

Samples per Map = 1

Wrote input for Map #0

Starting Job

．．．．．．

Job Finished in 15.972 seconds

Estimated value of Pi is 4.00000000000000000000

 

证明成功，至此完成！