在最后加入
export Hadoop_HOME=/usr/local/hadoop-1.0.4
export PATH=$HADOOP_HOME/bin:$PATH
5,配置/etc/hosts
192.168.1.11 namenode0
#192.168.1.12 namenode1
192.168.1.13 datanode00
192.168.1.14 datanode01
6,配置 conf/masters 和 conf/slaves
conf/masters
192.168.1.11
conf/slaves
192.168.1.13
192.168.1.14
7,配置 conf/hadoop-env.sh
加入
export JAVA_HOME=/usr/java/jdk1.6.0_21
8,配置 conf/core-site.xml
加入
<name>fs.default.name</name>
<!--<value>hdfs://0.0.0.0:9000</value> -->
<value>hdfs://namenode0:9000</value>
</property>
<property>
<name>hadoop.tmp.dir</name>
<value>/usr/local/hadoop/tmp</value>
</property>
9,配置 conf/hdfs-site.xml
加入
<property>
<name>dfs.replication</name>
<value>2</value>
</property>
<property>
<name>dfs.http.address</name>
<!-- <value>0.0.0.0:50070</value>backup -->
<value>192.168.1.11:50070</value>
</property>
<property>
<name>dfs.name.dir</name>
<value>/usr/local/hadoop/local/namenode</value>
</property>
<property>
<name>dfs.name.edits.dir</name>
<value>/usr/local/hadoop/local/editlog</value>
</property>
<property>
<name>dfs.data.dir</name>
<value>/usr/local/hadoop/block</value>
</property>
10,建立相关的目录（程序会根据上面的配置自动创建目录）
mkdir /usr/local/hadoop/tmp //hadoop临时目录
mkdir /usr/local/hadoop/local
mkdir /usr/local/hadoop/local/namenode //镜像存储目录
mkdir /usr/local/hadoop/local/editlog //日志存储目录
mkdir /usr/local/hadoop/block //数据块存储目录
如namenode00,虽然建立的block目录单，它不需要在目录，程序就不会创建，而datanode只会创建tmp和block，不会创建local。
注意如果在cygwin下这样配置，不会自动创建，而且手动创建后需要修改权限为755 ，不然会出现tasktracker启动失败问题，可以看tasktrack的日志
[user@namenode0 local]$ ls -sR hadoop
hadoop:
total 16
8 local 8 tmp
hadoop/local:
total 16
8 editlog 8 namenode
hadoop/local/editlog:
total 28
8 current 8 image 4 in_use.lock 8 previous.checkpoint
hadoop/local/editlog/current:
total 32
16 edits 8 fstime 8 VERSION
hadoop/local/editlog/image:
total 8
8 fsimage
hadoop/local/editlog/previous.checkpoint:
total 24
8 edits 8 fstime 8 VERSION
hadoop/local/namenode:
total 28
8 current 8 image 4 in_use.lock 8 previous.checkpoint
hadoop/local/namenode/current:
total 24
8 fsimage 8 fstime 8 VERSION
hadoop/local/namenode/image:
total 8
8 fsimage
hadoop/local/namenode/previous.checkpoint:
total 24
8 fsimage 8 fstime 8 VERSION
hadoop/tmp:
total 8
8 dfs
hadoop/tmp/dfs:
total 8
8 namesecondary
hadoop/tmp/dfs/namesecondary:
total 4
4 in_use.lock
11.修改namenode0配置文件masters
12 修改namenode0配置文件slaves
13.修改mapred-site.xml
<property>
<name>mapred.job.tracker</name>
<value>namenode0:9001</value>
</property>
<property>
<name>mapred.child.tmp</name>
<value>/usr/local/hadoop/tmp</value>
</property>
14，将hadoop远程copy到其他节点
scp -r hadoop-1.0.4 192.168.1.13:/usr/local/) 192.168.1.12也可以写为datanode00 （也可scp -r hadoop-1.0.4user@192.168.1.13:/usr/local/）
15，格式化Active namenode（192.168.1.11）
bin/hadoop namenode -format
16，启动集群 bin/start-all.sh