hadoop 2.2.0安装

hadoop 2.2.0相比于hadoop 0.20.X的安装很不一样。

1、准备三台机器 test01(master)，test02(slaves), test03(slaves)，并把test01的公钥放到test02/test03的~/.ssh/authorized_keys以打通master和集群的通信。

2、下载hadoop-2.2.0.tar.gz

3、在hadoop 2.2.0进行配置，配置完成后再拷贝到其他slaves上

4、配置文件都在$HADOOP_HOME/etc/hadoop下

a、hadoop-env.sh:

 

替换exportJAVA_HOME=${JAVA_HOME}为如下：

export JAVA_HOME=Java安装路径

b、core-site.xml:

 

<configuration>
<property>
   <name>fs.default.name</name>
   <value>hdfs://test01:9000</value> 
</property>
<property>
   <name>hadoop.tmp.dirname</name>
   <value>/root/hadoop/data/</value>
</property>
</configuration>

 

注意：

test01的master的ip。

要事先创建好/root/hadoop/data

c、mapred-site.xml:

 

<configuration>
<property>
        <name>mapred.job.tracker</name>
        <value>test01:9001</value>
   </property>
</configuration>

d、hdfs-site.xml:

 

 

<configuration>
<property>
  <name>dfs.replication</name>
  <value>2</value>
  <description>Default block replication.
  The actual number of replications can be specified when the file is created.
  The default is used if replication is not specified in create time.
  </description>
</property>
</configuration>

e、slaves:

 

 

test02
test03

5、配置完成后，将test01的hadoop文件夹拷到test02和test03上

 

6、format

在test01上运行$HADOOP_HOME/bin/hadoop namenode -format

如果执行成功，你会在日志中(倒数几行)找到如下成功的提示信息：

 

common.Storage: Storage directory /home/hduser/hadoop/tmp/hadoop-hduser/dfs/namehas been successfully formatted.

注意如果在test02又跑了一次format，运行时会报

Incompatible namespaceIDs in /tmp/hadoop-root/dfs/data

的错误，导致DataNode启动不起来，解决方法是删除集群中各台机器的rm -rf /tmp/hadoop-root/* ，然后重新format就行

 

7、启动hadoop

$HADOOP_HOME/sbin/start-dfs.sh

$HADOOP_HOME/sbin/start-yarn.sh

note:这个版本已经不推荐使用start-all.sh命令

成功完成后：

test01 jps:

 

11047 ResourceManager
10906 SecondaryNameNode
11966 Jps
10724 NameNode

test02/test03 jps:

 

 

6406 NodeManager
6288 DataNode
6508 Jps

 

使用$HADOOP_HOME/bin/hdfs dfsadmin -report查看hdfs状态：

 

Configured Capacity: 52840488960 (49.21 GB)
Present Capacity: 45652713472 (42.52 GB)
DFS Remaining: 45652664320 (42.52 GB)
DFS Used: 49152 (48 KB)
DFS Used%: 0.00%
Under replicated blocks: 0
Blocks with corrupt replicas: 0
Missing blocks: 0

-------------------------------------------------
Datanodes available: 2 (2 total, 0 dead)

Live datanodes:
Name: 192.168.2.106:50010 (test02)
Hostname: test02
Decommission Status : Normal
Configured Capacity: 26420244480 (24.61 GB)
DFS Used: 24576 (24 KB)
Non DFS Used: 3575242752 (3.33 GB)
DFS Remaining: 22844977152 (21.28 GB)
DFS Used%: 0.00%
DFS Remaining%: 86.47%
Last contact: Sat Nov 16 12:39:27 CST 2013


Name: 192.168.2.99:50010 (test03)
Hostname: test03
Decommission Status : Normal
Configured Capacity: 26420244480 (24.61 GB)
DFS Used: 24576 (24 KB)
Non DFS Used: 3612532736 (3.36 GB)
DFS Remaining: 22807687168 (21.24 GB)
DFS Used%: 0.00%
DFS Remaining%: 86.33%
Last contact: Sat Nov 16 12:39:27 CST 2013

8、关闭命令

 

stop-yarn.shstop-dfs.sh

9、hadoop相关dfs命令已经和以前不一样，如hadoop dfs -ls /tmp 要修改成$HADOOP_HOME/bin/hdfs dfs -ls /tmp

10、查看管理页面： http://test01:50070/

 

遇到的问题：

1、本地库问题

WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable