Linux下部署Hadoop伪分布模式

Hadoop版本为1.2.1

Distribution为Fedora19并使用hadoop账号安装

第一步：配置ssh本地登录证书（虽然为伪分布模式，Hadoop依然会使用SSH进行通信）

[hadoop@promote ~]$ which ssh
/usr/bin/ssh
[hadoop@promote ~]$ which ssh-keygen
/usr/bin/ssh-keygen
[hadoop@promote ~]$ which sshd
/usr/sbin/sshd
[hadoop@promote ~]$ ssh-keygen -t rsa

然后一路回车

Generating public/private rsa key pair.
Enter file in which to save the key (/home/hadoop/.ssh/id_rsa): 
Created directory '/home/hadoop/.ssh'.
Enter passphrase (empty for no passphrase): 
Enter same passphrase again: 
Passphrases do not match.  Try again.
Enter passphrase (empty for no passphrase): 
Enter same passphrase again: 
Your identification has been saved in /home/hadoop/.ssh/id_rsa.
Your public key has been saved in /home/hadoop/.ssh/id_rsa.pub.
The key fingerprint is:
2f:a9:60:c7:dc:38:8f:c7:bb:70:de:d4:39:c3:39:87 hadoop@promote.cache-dns.local
The key's randomart image is:
+--[ RSA 2048]----+
|                 |
|                 |
|                 |
|                 |
|        S        |
|     o o o o +   |
|    o B.= o E .  |
|   . o Oo+   =   |
|      o.=o.      |
+-----------------+

最终将在/home/hadoop/.ssh/路径下生成私钥id_rsa和公钥id_rsa.pub

[hadoop@promote .ssh]$ cd /home/hadoop/.ssh/
[hadoop@promote .ssh]$ ls
id_rsa  id_rsa.pub

修改sshd服务配置文件：

[hadoop@promote .ssh]$ su root
密码：
[root@promote .ssh]# vi /etc/ssh/sshd_config 

启用RSA加密算法验证

RSAAuthentication yes
PubkeyAuthentication yes# The default is to check both .ssh/authorized_keys and .ssh/authorized_keys2
# but this is overridden so installations will only check .ssh/authorized_keys
AuthorizedKeysFile      .ssh/authorized_keys

保存并退出，然后重启sshd服务

[root@promote .ssh]# service sshd restart
Redirecting to /bin/systemctl restart  sshd.service

然后切换回hadoop用户，将ssh证书公钥拷贝至/home/hadoop/.ssh/authorized_keys文件中

[root@promote .ssh]# su hadoop
[hadoop@promote .ssh]$ cat id_rsa.pub >> authorized_keys

修改~/.ssh/authorized_keys文件的权限为644，~/.ssh文件夹的权限为700，/home/hadoop文件夹的权限为700（权限正确是成功认证的先决条件）

[hadoop@promote .ssh]$ chmod 644 authorized_keys 
[hadoop@promote .ssh]$ ssh 192.168.211.129
The authenticity of host 192.168.211.129(192.168.211.129)' can't be established.
RSA key fingerprint is 25:1f:be:72:7b:83:8e:c7:96:b6:71:35:fc:5d:2e:7d.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added '192.168.211.129' (RSA) to the list of known hosts.
Last login: Thu Feb 13 23:42:43 2014 

第一次登陆将会将证书内容保存在/home/hadoop/.ssh/known_hosts文件中，以后再次登陆将不需要输入密码

[hadoop@promote .ssh]$ ssh 192.168.211.129
Last login: Thu Feb 13 23:46:04 2014 from 192.168.211.129

至此ssh证书部分配置完成

第二步：安装JDK

[hadoop@promote ~]$ java -version
java version "1.7.0_25"
OpenJDK Runtime Environment (fedora-2.3.10.3.fc19-i386)
OpenJDK Client VM (build 23.7-b01, mixed mode)

将OpenJDK换为Oracle的Java SE

[hadoop@promote .ssh]$ cd ~
[hadoop@promote ~]$ uname -i
i386

在Oracle的官网下载jdk-6u45-linux-i586.bin后上传至服务器，赋予权限并进行安装，最后删除安装包

[hadoop@promote ~]$ chmod u+x jdk-6u45-linux-i586.bin 
[hadoop@promote ~]$ ./jdk-6u45-linux-i586.bin 
[hadoop@promote ~]$ rm -rf jdk-6u45-linux-i586.bin 

出现以下结果说明JDK成功安装：

[hadoop@promote ~]$ /home/hadoop/jdk1.6.0_45/bin/java -version
java version "1.6.0_45"
Java(TM) SE Runtime Environment (build 1.6.0_45-b06)
Java HotSpot(TM) Client VM (build 20.45-b01, mixed mode, sharing)

第三步：安装Hadoop

在Hadoop官网下载hadoop-1.2.1.tar.gz并上传至服务器/home/hadoop路径下

[hadoop@promote ~]$ tar -xzf hadoop-1.2.1.tar.gz
[hadoop@promote ~]$ rm -rf hadoop-1.2.1.tar.gz 
[hadoop@promote ~]$ cd hadoop-1.2.1/conf/
[hadoop@promote conf]$ vi hadoop-env.sh

将JAVA_HOME指向第二步安装的JDK所在目录

# The java implementation to use.  Required.
export JAVA_HOME=/home/hadoop/jdk1.6.0_45

保存并退出

[hadoop@promote ~]$ vi ~/.bash_profile 

环境变量PATH后接上Hadoop和JDK的bin目录

......
PATH=$PATH:$HOME/.local/bin:$HOME/bin:/home/hadoop/hadoop-1.2.1/bin:/home/hadoop/jdk1.6.0_45/bin
export PATH

保存并退出，退出登录并重新使用hadoop账号登录，如果出现如下结果，说明环境变量PATH设置成功

[hadoop@promote ~]$ echo $PATH
/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin:/home/hadoop/.local/bin:
/home/hadoop/bin:/home/hadoop/hadoop-1.2.1/bin:/home/hadoop/jdk1.6.0_45/bin

第四步：修改Hadoop配置文件

修改core-site.xml（使用IP地址而不是主机名或localhost的好处是不需要修改/etc/hosts）：

<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?><configuration>
<property>
<name>fs.default.name</name>
<value>hdfs://192.168.211.129:9000</value>
</property>
<property>
<name>hadoop.tmp.dir</name>
<value>/home/hadoop/hadooptmp</value>
</property>
</configuration>

修改mapred-site.xml：

<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?><configuration>
<property>
<name>mapred.job.tracker</name>
<value>192.168.211.129:9001</value>
</property>
</configuration>

修改hdfs-site.xml：

<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?><configuration>
<property>
<name>dfs.replication</name>
<value>1</value>
</property>
</configuration>

master中指定的SNN节点和slaves中指定的从节点位置均为本地

[hadoop@promote conf]$ cat masters 
192.168.211.129
[hadoop@promote conf]$ cat slaves 
192.168.211.129

第五步：初始化HDFS文件系统

特别要注意的是，Hadoop并不识别带“_”的主机名，所以如果你的主机名带有“_”，一定要进行修改，修改方式参照http://blog.csdn.net/a19881029/article/details/20485079

[hadoop@fedora ~]$ hadoop namenode -format
14/03/04 22:13:41 INFO namenode.NameNode: STARTUP_MSG: 
/************************************************************
STARTUP_MSG: Starting NameNode
STARTUP_MSG:   host = fedora/127.0.0.1
STARTUP_MSG:   args = [-format]
STARTUP_MSG:   version = 1.2.1
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
STARTUP_MSG:   java = 1.6.0_45
************************************************************/
14/03/04 22:13:42 INFO util.GSet: Computing capacity for map BlocksMap
14/03/04 22:13:42 INFO util.GSet: VM type       = 32-bit
14/03/04 22:13:42 INFO util.GSet: 2.0% max memory = 1013645312
14/03/04 22:13:42 INFO util.GSet: capacity      = 2^22 = 4194304 entries
14/03/04 22:13:42 INFO util.GSet: recommended=4194304, actual=4194304
14/03/04 22:13:42 INFO namenode.FSNamesystem: fsOwner=hadoop
14/03/04 22:13:42 INFO namenode.FSNamesystem: supergroup=supergroup
14/03/04 22:13:42 INFO namenode.FSNamesystem: isPermissionEnabled=true
14/03/04 22:13:42 INFO namenode.FSNamesystem: dfs.block.invalidate.limit=100
14/03/04 22:13:42 INFO namenode.FSNamesystem: isAccessTokenEnabled=false accessKeyUpdateInterval=0 min(s), accessTokenLifetime=0 min(s)
14/03/04 22:13:42 INFO namenode.FSEditLog: dfs.namenode.edits.toleration.length = 0
14/03/04 22:13:42 INFO namenode.NameNode: Caching file names occuring more than 10 times
14/03/04 22:13:53 INFO common.Storage: Image file /tmp/hadoop-hadoop/dfs/name/ current/fsimage of size 112 bytes saved in 0 seconds.
14/03/04 22:13:53 INFO namenode.FSEditLog: closing edit log: position=4, editlog=/tmp/hadoop-hadoop/dfs/name/current/edits
14/03/04 22:13:53 INFO namenode.FSEditLog: close success: truncate to 4, editlog=/tmp/hadoop-hadoop/dfs/name/current/edits
14/03/04 22:13:53 INFO common.Storage: Storage directory /tmp/hadoop-hadoop/dfs/name has been successfully formatted.
14/03/04 22:13:53 INFO namenode.NameNode: SHUTDOWN_MSG: 
/************************************************************
SHUTDOWN_MSG: Shutting down NameNode at fedora/127.0.0.1
************************************************************/

第六步：启动Hadoop

[hadoop@fedora logs]$ start-all.sh 
starting namenode, logging to /home/hadoop/hadoop-1.2.1/libexec/../logs/hadoop-hadoop-namenode-fedora.out
localhost: starting datanode, logging to /home/hadoop/hadoop-1.2.1/libexec/../logs/hadoop-hadoop-datanode-fedora.out
localhost: starting secondarynamenode, logging to /home/hadoop/hadoop-1.2.1/libexec/../logs/hadoop-hadoop-secondarynamenode-fedora.out
starting jobtracker, logging to /home/hadoop/hadoop-1.2.1/libexec/../logs/hadoop-hadoop-jobtracker-fedora.out
localhost: starting tasktracker, logging to /home/hadoop/hadoop-1.2.1/libexec/../logs/hadoop-hadoop-tasktracker-fedora.out
hadoop@fedora logs]$ jps
2099 SecondaryNameNode
2184 JobTracker
1976 DataNode
2365 Jps
1877 NameNode
2289 TaskTracker

可以看到所有Hadoop守护进程均已启动

再看看日志文件中有没有报错，如果没有报错，说明Hadoop已经启动成功了

[hadoop@fedora hadoop]$ hadoop dfsadmin -report
Configured Capacity: 39474135040 (36.76 GB)
Present Capacity: 33661652992 (31.35 GB)
DFS Remaining: 33661612032 (31.35 GB)
DFS Used: 40960 (40 KB)
DFS Used%: 0%
Under replicated blocks: 0
Blocks with corrupt replicas: 0
Missing blocks: 0-------------------------------------------------
Datanodes available: 1 (1 total, 0 dead)Name: 192.168.211.129:50010
Decommission Status : Normal
Configured Capacity: 39474135040 (36.76 GB)
DFS Used: 40960 (40 KB)
Non DFS Used: 5812482048 (5.41 GB)
DFS Remaining: 33661612032(31.35 GB)
DFS Used%: 0%
DFS Remaining%: 85.28%
Last contact: Thu Mar 06 09:48:17 CST 2014

尝试执行Map/Reduce任务时也是没问题的