Hadoop1.2.1伪分布模式安装指南

一、前置条件

（1）Linux可以用作开发平台及产品平台。

2、安装jdk 1.6或以上

$ ssh-keygen -t dsa -P ‘‘ -f ~/.ssh/id_dsa 
$ cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys

二、基本准备

[plain] view plaincopy

1. [[email protected] jediael]$wget http://mirror.bit.edu.cn/apache/hadoop/common/hadoop-1.2.1/hadoop-1.2.1.tar.gz
2. [email protected] jediael]$ tar -zxvf hadoop-1.2.1-bin.tar.gz

2、修改conf/hadoop-env.sh，添加JAVA_HOME变量

增加JAVA_HOME

• export JAVA_HOME=/usr/java/jdk1.7.0_51

[plain] view plaincopy

1. [[email protected] hadoop-1.2.1]$ bin/hadoop
2. Usage: hadoop [--config confdir] COMMAND
3. where COMMAND is one of:
4. namenode -format format the DFS filesystem
5. secondarynamenode run the DFS secondary namenode
6. namenode run the DFS namenode
7. datanode run a DFS datanode
8. dfsadmin run a DFS admin client
9. mradmin run a Map-Reduce admin client
10. fsck run a DFS filesystem checking utility
11. fs run a generic filesystem user client

三、配置伪分布模式

<configuration>
     <property>
         <name>fs.default.name</name>
         <value>hdfs://localhost:9000</value>
     </property>
</configuration>

<configuration>
     <property>
         <name>dfs.replication</name>
         <value>1</value>
     </property>
</configuration>

3、配置mapred-site.xml，增加以下属性

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

四、启动hadoop

[plain] view plaincopy

1. [[email protected] hadoop-1.2.1]$ bin/hadoop namenode -format
2. 14/08/16 23:50:02 INFO namenode.NameNode: STARTUP_MSG:
3. /************************************************************
4. STARTUP_MSG: Starting NameNode
5. STARTUP_MSG: host = jediael/10.171.29.191
6. STARTUP_MSG: args = [-format]
7. STARTUP_MSG: version = 1.2.1
8. STARTUP_MSG: build = https://svn.apache.org/repos/asf/hadoop/common/branches/branch-1.2 -r 1503152; compiled by ‘mattf‘ on Mon Jul 22 15:23:09 PDT 2013
9. STARTUP_MSG: java = 1.7.0_51
10. ************************************************************/
11. 14/08/16 23:50:02 INFO util.GSet: Computing capacity for map BlocksMap
12. 14/08/16 23:50:02 INFO util.GSet: VM type = 64-bit
13. 14/08/16 23:50:02 INFO util.GSet: 2.0% max memory = 1013645312
14. 14/08/16 23:50:02 INFO util.GSet: capacity = 2^21 = 2097152 entries
15. 14/08/16 23:50:02 INFO util.GSet: recommended=2097152, actual=2097152
16. 14/08/16 23:50:02 INFO namenode.FSNamesystem: fsOwner=jediael
17. 14/08/16 23:50:02 INFO namenode.FSNamesystem: supergroup=supergroup
18. 14/08/16 23:50:02 INFO namenode.FSNamesystem: isPermissionEnabled=true
19. 14/08/16 23:50:02 INFO namenode.FSNamesystem: dfs.block.invalidate.limit=100
20. 14/08/16 23:50:02 INFO namenode.FSNamesystem: isAccessTokenEnabled=false accessKeyUpdateInterval=0 min(s), accessTokenLifetime=0 min(s)
21. 14/08/16 23:50:02 INFO namenode.FSEditLog: dfs.namenode.edits.toleration.length = 0
22. 14/08/16 23:50:02 INFO namenode.NameNode: Caching file names occuring more than 10 times
23. 14/08/16 23:50:03 INFO common.Storage: Image file /tmp/hadoop-jediael/dfs/name/current/fsimage of size 113 bytes saved in 0 seconds.
24. 14/08/16 23:50:03 INFO namenode.FSEditLog: closing edit log: position=4, editlog=/tmp/hadoop-jediael/dfs/name/current/edits
25. 14/08/16 23:50:03 INFO namenode.FSEditLog: close success: truncate to 4, editlog=/tmp/hadoop-jediael/dfs/name/current/edits
26. 14/08/16 23:50:03 INFO common.Storage: Storage directory /tmp/hadoop-jediael/dfs/name has been successfully formatted.
27. 14/08/16 23:50:03 INFO namenode.NameNode: SHUTDOWN_MSG:
28. /************************************************************
29. SHUTDOWN_MSG: Shutting down NameNode at jediael/10.171.29.191
30. ************************************************************/

[plain] view plaincopy

1. [[email protected] hadoop-1.2.1]# bin/start-all.sh
2. starting namenode, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-root-namenode-jediael.out
3. localhost: starting datanode, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-root-datanode-jediael.out
4. localhost: starting secondarynamenode, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-root-secondarynamenode-jediael.out
5. starting jobtracker, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-root-jobtracker-jediael.out
6. localhost: starting tasktracker, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-root-tasktracker-jediael.out

如何设置普通用户的ssh免密码用户登录？同样方法不成功，sudo也不成功。待解决。

默认情况下，日志将被输出至{HADOOP_HOME}/logs，除非修改了${HADOOP_LOG_DIR}。

3、访问以下2个页面，验证是否已经安装成功

• NameNode - http://localhost:50070/
• JobTracker - http://localhost:50030/

[plain] view plaincopy

1. [[email protected] hadoop-1.2.0]# jps
2. 3148 JobTracker
3. 3280 TaskTracker
4. 3052 SecondaryNameNode
5. 2920 DataNode
6. 2801 NameNode
7. 3442 Jps

五、使用一个简单的hadoop程序验证环境

Hadoop1.2.1伪分布模式安装指南,布布扣,bubuko.com