1.下载windows环境下编译的hadoop-2.7.2.x64win.zip
2.解压至D:\BigData\hadoop-2.7.2
3.修改D:\BigData\hadoop-2.7.2\etc\hadoop\hadoop-env.cmd
set HADOOP_PREFIX=D:\bigdata\hadoop-2.7.2
4.初始化环境变量
D:\BigData\hadoop-2.7.2\etc\hadoop>hadoop-env.cmd
5.格式化namenode
D:\BigData\hadoop-2.7.2\bin>hdfs namenode -format
成功会出现：successfully formatted
6.启动namenode和datanode
D:\BigData\hadoop-2.7.2\sbin>start-dfs.cmd
查看是否启动成功，在原先窗口输入jps查看,如图：
测试：运行hdfs命令，上传一个文件，在bin目录下创建myfile.txt文件
执行：D:\BigData\hadoop-2.7.2\bin>hdfs dfs -put myfile.txt /
查看上传的文件信息，执行：
D:\BigData\hadoop-2.7.2\bin>hdfs dfs -ls /
7.启动yarn
D:\BigData\hadoop-2.7.2\sbin>start-yarn.cmd
输入jps命令可查看当前启动的节点
8.Web端查看hadoop作业
http://localhost:8088/cluster
9.结束hadoop
D:\BigData\hadoop-2.7.2\sbin>stop-yarn.cmd
D:\BigData\hadoop-2.7.2\sbin>stop-dfs.cmd
参考网址：http://blog.csdn.net/dream_an/article/details/51862640