hadoop伪分布环境搭建
Hadoop伪分布环境搭建(网易云课堂hadoop大数据实战学习笔记)
工具:VMware WorkStation 8.0.2;
CentOS6.4;
hadoop-1.1.2.tar.gz;
Xshell 4,作为登录工具,可以不启动CentOS的GUI(实验机的配置较低,可以省资源),“远程登录”CentOS;
1、root登录,安装jdk,直接看结果,安装过程和环境变量配置不再细说;
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175353433-1162847653.png)
2、设置SSH免密码登录;
(1)关闭防火墙;
service iptables stop 或者 setup
查看状态:service iptables status
(2)设置DNS解析
vim /etc/hosts ip 主机名(这里出错了,后面有解释)
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175353823-1137426996.png)
(3)vim /etc/selinux/config
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175354308-1029736014.png)
ssh-keygen -t dsa 一直回车;
cd ~/.ssh
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175354573-134294967.png)
cat id_dsa.pub >> authorized_keys
测试是否成功:ssh localhost,没有输密码,免密码登录
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175354933-1421934187.png)
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175355229-752772542.png)
3、安装hadoop;
cd /usr/local;
tar -xzvf hadoop-1.1.2.tar.gz ;
mv hadoop-1.1.2 hadoop;
配置Hadoop的环境变量:
export HADOOP_HOME=/usr/local/hadoop
export PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175355604-1078069050.png)
修改配置文件:
(1)hadoop-env.sh
更改java_home 变量
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175355901-687080617.png)
(2)core-site.xml
1 <configuration> 2 <property> 3 <name>fs.default.name</name> 4 <value>hdfs://hadoop0:9000</value> 5 </property> 6 <property> 7 <name>hadoop.tmp.dir</name> 8 <value>/usr/local/hadoop/hadoop_${user.name}</value> 9 </property> 10 </configuration>
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175356386-1647954190.png)
(3)hdfs-site.xml
1 <configuration> 2 2 <property> 3 3 <name>dfs.replication</name> 4 4 <value>1</value> //实验环境,备份设置为1 5 5 </property> 6 6 <property> 7 7 <name>dfs.permissions</name> 8 8 <value>false</value> //实验环境,不设置权限 9 9 </property> 10 10 </configuration>
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175356823-1622038599.png)
(4)vim mapred-site.xml
1 <configuration> 2 <property> 3 <name>mapred.job.tracker</name> 4 <value>hadoop0:9001</value> 5 </property> 6 </configuration>
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175357261-801686970.png)
hadoop namenode -format
格式化时候,出现错误:
java.net.UnknownHostException: hostname: hostname: Name or service not known
查看之后是犯了一个低级错误,DNS解析时候;主机名是CenOS(装系统时候设定的),但是/etc/hosts文件中写的是hadoop0,
因此,更改了一下主机名为hadoop0,修改/etc/sysconfig/network中的hostname为hadoop0,重启系统,再次格式化,正常;
启动hadoop,命令:start-all.sh
查看相关的进程: jps
namenode、datanode、jobtracker、tasktracker、secondaryNamenode
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175357573-1394193315.png)
查看50030端口:
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175358464-748445992.png)
查看50070端口:
![](https://images2015.cnblogs.com/blog/921208/201603/921208-20160324175359542-892357377.png)
至此,说明hadoop伪分布式环境搭建完成。