原文来自林羽飞扬 ubuntu版本12.04.3 64位 hadoop是用运行在java虚拟机上的，所以需要先安装jdk，jdk安装配置方法在另一篇博文ubuntu12.04下安装jdk1.7 源码包准备： 我下载的是hadoop-1.2.1.tar.gz,这个版本比较稳定，下载地址可以到官网提供的镜像http://w
原文来自林羽飞扬
ubuntu版本12.04.3 64位
hadoop是用运行在java虚拟机上的，所以需要先安装jdk，jdk安装配置方法在另一篇博文ubuntu12.04下安装jdk1.7
源码包准备：
我下载的是hadoop-1.2.1.tar.gz,这个版本比较稳定，下载地址可以到官网提供的镜像http://www.apache.org/dyn/closer.cgi/hadoop/common/
1、创建hadoop管理员帐号
直接在终端执行如下命令行：
<span>1</span> <span>sudo</span> adduser hadoop
然后要求你设置hadoop帐户密码，这个命令是添加一个名为hadoop的标准帐户，我们需要的是管理员帐号
可以直接在图形界面下修改hadoop权限，将鼠标点击右上角的一个人头处，浮现列表，点击“用户账户”，解锁，然后更改为管理员权限
2、安装ssh服务
ssh可以实现远程登录和管理，详细情况请google百度
ubuntu默认并没有安装ssh服务，如果通过ssh链接ubuntu，需要自己手动安装ssh-server。命令行：
<span>1</span> <span>sudo</span> apt-get <span>install</span> <span>ssh</span> openssh-server
3、ssh无密码验证登录
创建ssh-key，这里我们采用rsa方式，命令行如下：
<span>1</span> <span>ssh-keygen</span> -t rsa -P <span>""</span>
出现一个图形，出现的图形就是密码，不用管它
<span>1</span> <span>cat</span> ~/.<span>ssh</span>/id_rsa.pub >> authorized_keys
然后即可无密码验证登录了，如下：
<span>1</span> <span>ssh</span> localhost
退出命令行为：
exit
4、解压hadoop源码包
终端下进入hadoop源码包所在目录，使用复制命令把hadoop源码包复制到/home/hadoop下
<span>1</span> <span>cp</span> hadoop-<span>1.2</span>.<span>1</span>.<span>tar</span>.gz /home/hadoop
然后解压,命令行如下
<span>tar</span> -xzvf *.tag.gz
5、配置hadoop的hadoop/conf下的hadoop-env.sh，core-site.xml，mapred-site.xml，hdfs-site.xml
配置hadoop-1.2.1/conf/hadoop-env.sh，命令行：
<span>1</span> gedit /home/hadoop/hadoop-<span>1.2</span>.<span>1</span>/conf/hadoop-<span>env</span>.<span>sh</span>
ctrl + f 搜索到JAVA_HOME
把前面的#去掉，加上本系统jdk路径，保存退出
配置hadoop-1.2.1/conf/core-site.xml，命令行：
gedit /home/hadoop/hadoop-<span>1.2</span>.<span>1</span>/conf/core-site.xml
在hadoop新建hadoop_tmp目录，原因是http://blog.csdn.net/bychjzh/article/details/7830508
将如下
<span><?</span><span>xml version="1.0"</span><span>?></span>
<span><?</span><span>xml-stylesheet type="text/xsl" href="configuration.xsl"</span><span>?></span>
<span><!--</span><span> Put site-specific property overrides in this file. </span><span>--></span>
<span><</span><span>configuration</span><span>></span>
<span><</span><span>property</span><span>></span>
<span><</span><span>name</span><span>></span>fs.default.name<span></</span><span>name</span><span>></span>
<span><</span><span>value</span><span>></span>hdfs://localhost:9000<span></</span><span>value</span><span>></span>
<span></</span><span>property</span><span>></span>
<span></</span><span>configuration</span><span>></span>
配置hadoop-1.2.1/conf/mapre-site.xml，命令行：
<span>1</span> gedit /home/hadoop/hadoop-<span>1.2</span>.<span>1</span>/conf/mapred-site.xml
将如下
<span><?</span><span>xml version="1.0"</span><span>?></span>
<span><?</span><span>xml-stylesheet type="text/xsl" href="configuration.xsl"</span><span>?></span>
<span><!--</span><span> Put site-specific property overrides in this file. </span><span>--></span>
<span><</span><span>configuration</span><span>></span>
<span><</span><span>property</span><span>></span>
<span><</span><span>name</span><span>></span>mapred.job.tracker<span></</span><span>name</span><span>></span>
<span><</span><span>value</span><span>></span>localhost:9001<span></</span><span>value</span><span>></span>
<span></</span><span>property</span><span>></span>
<span></</span><span>configuration</span><span>></span>
配置hadoop-1.2.1/conf/hdfs-site.xml，命令行：
<span>1</span> gedit /home/hadoop/hadoop-<span>1.2</span>.<span>1</span>/conf/hdfs-site.xml
将如下
<span><?</span><span>xml version="1.0"</span><span>?></span>
<span><?</span><span>xml-stylesheet type="text/xsl" href="configuration.xsl"</span><span>?></span>
<span><!--</span><span> Put site-specific property overrides in this file. </span><span>--></span>
<span><</span><span>configuration</span><span>></span>
<span><</span><span>property</span><span>></span>
<span><</span><span>name</span><span>></span>dfs.replication<span></</span><span>name</span><span>></span>
<span><</span><span>value</span><span>></span>1<span></</span><span>value</span><span>></span>
<span></</span><span>property</span><span>></span>
<span></</span><span>configuration</span><span>></span>
至此hadoop的安装配置已经完毕，稍后的是hadoop的初次运行操作
6、格式化hdfs文件系统
进入hadoop-1.2.1
/bin/hadoop namenode -format
当你看到下图时
恭喜你，hdsf文件系统格式化成功
7、启动hadoop服务
/bin/start-all.<span>sh</span>
出现如下画面
jps
jps是查看java虚拟机运行的java线程
然后出现如下画面
不计jps，有五个hadoop相关线程，恭喜你，hadoop安装配置成功，运行正常。
然后可以退出hadoop。，以后再用时再启动，导入数据
<span>1</span> bin/stop-all.<span>sh</span>
