hadoop的Linux下简单安装步骤

13-08-26 banq
首先安装Hadoop cluster版本:
$ curl -O http://archive.cloudera.com/cdh/3/hadoop-0.20.2-cdh3u6.tar.gz
$ cd /usr/local
$ tar xzvf /path/to/hadoop-0.20.2-cdh3u6.tar.gz
$ ln -s hadoop-0.20.2-cdh3u6 hadoop

配置工作:
$ cd hadoop
$ mv conf conf.original
$ cp -Rp example-confs/conf.pseudo conf


将/usr/local/hadoop/bin加入PATH环境
$ export PATH=/usr/local/hadoop/bin:$PATH

格式化HDFS
建立临时文件目录 /var/lib/hadoop-0.20/
chmod 0777 -R /var/lib/hadoop-0.20/
chmod 0777 -R /usr/local/hadoop/
以普通用户登录,非ROOT用户,然后运行:
hadoop namenode -format

出现:O common.Storage: Storage directory /var/lib/hadoop-0.20/cache/hadoop/dfs/name has been successfully formatted.
表示OK。

分别运行:
$ hadoop namenode
$ hadoop datanode
$ hadoop jobtracker
$ hadoop tasktracker