安装hadoop1.2.1集群环境

一、规划
（一）硬件资源
10.171.29.191 master
10.173.54.84  slave1
10.171.114.223 slave2

（二）基本资料
用户：  jediael
目录：/opt/jediael/

二、环境配置
（一）统一用户名密码，并为jediael赋予执行所有命令的权限

#passwd
# useradd jediael
# passwd jediael
# vi /etc/sudoers

增加以下一行：

jediael ALL=(ALL) ALL

（二）创建目录/opt/jediael

$sudo chown jediael:jediael /opt
$ cd /opt
$ sudo mkdir jediael

注意:/opt必须是jediael的，否则会在format namenode时出错。

（三）修改用户名及/etc/hosts文件
1、修改/etc/sysconfig/network

NETWORKING=yes
HOSTNAME=*******

2、修改/etc/hosts

10.171.29.191 master
10.173.54.84  slave1
10.171.114.223 slave2

注 意hosts文件不能有127.0.0.1  *****配置，否则会导致出现异常。org.apache.hadoop.ipc.Client: Retrying connect to server: master/10.171.29.191:9000. Already trie
3、hostname命令

hostname ****

（四）配置免密码登录
以上命令在master上使用jediael用户执行：

$ ssh-keygen -t dsa -P '' -f ~/.ssh/id_dsa
$ cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys

然后，将authorized_keys复制到slave1,slave2

scp ~/.ssh/authorized_keys slave1:~/.ssh/
scp ~/.ssh/authorized_keys slave2:~/.ssh/

注意
（1）若提示.ssh目录不存在，则表示此机器从未运行过ssh，因此运行一次即可创建.ssh目录。
（2）.ssh/的权限为600,authorized_keys的权限为700，权限大了小了都不行。


（五）在3台机器上分别安装java，并设置相关环境变量
参考http://blog.csdn.net/jediael_lu/article/details/38925871

（六）下载hadoop-1.2.1.tar.gz，并将其解压到/opt/jediael

三、修改配置文件
【3台机器上均要执行】
（一）修改conf/hadoop_env.sh

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

（二）修改core-site.xml

<property>
 <name>fs.default.name</name>
 <value>hdfs://master:9000</value>
</property>
 
<property>
 <name>hadoop.tmp.dir</name>
 <value>/opt/tmphadoop</value>
</property> 

（三）修改hdfs-site.xml

<property>
 <name>dfs.replication</name>
 <value>2</value>
</property>

（四）修改mapred-site.xml

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

（五）修改master及slaves

master:
master

slaves:
slave1
slave2

可以在master中完成上述配置，然后使用scp命令复制到slave1与slave2上。
 如：

$scp core-site.xml slave2:/opt/jediael/hadoop-1.2.1/conf

四、启动并验证


1、格式 化namenode【此步骤在3台机器上均要运行】

[jediael@master hadoop-1.2.1]$  bin/hadoop namenode -format

15/01/21 15:13:40 INFO namenode.NameNode: STARTUP_MSG:
/************************************************************
STARTUP_MSG: Starting NameNode
STARTUP_MSG:   host = master/10.171.29.191
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 = 1.7.0_51
************************************************************/
Re-format filesystem in /opt/tmphadoop/dfs/name ? (Y or N) Y
15/01/21 15:13:43 INFO util.GSet: Computing capacity for map BlocksMap
15/01/21 15:13:43 INFO util.GSet: VM type       = 64-bit
15/01/21 15:13:43 INFO util.GSet: 2.0% max memory = 1013645312
15/01/21 15:13:43 INFO util.GSet: capacity      = 2^21 = 2097152 entries
15/01/21 15:13:43 INFO util.GSet: recommended=2097152, actual=2097152
15/01/21 15:13:43 INFO namenode.FSNamesystem: fsOwner=jediael
15/01/21 15:13:43 INFO namenode.FSNamesystem: supergroup=supergroup
15/01/21 15:13:43 INFO namenode.FSNamesystem: isPermissionEnabled=true
15/01/21 15:13:43 INFO namenode.FSNamesystem: dfs.block.invalidate.limit=100
15/01/21 15:13:43 INFO namenode.FSNamesystem: isAccessTokenEnabled=false accessKeyUpdateInterval=0 min(s), accessTokenLifetime=0 min(s)
15/01/21 15:13:43 INFO namenode.FSEditLog: dfs.namenode.edits.toleration.length = 0
15/01/21 15:13:43 INFO namenode.NameNode: Caching file names occuring more than 10 times
15/01/21 15:13:44 INFO common.Storage: Image file /opt/tmphadoop/dfs/name/current/fsimage of size 113 bytes saved in 0 seconds.
15/01/21 15:13:44 INFO namenode.FSEditLog: closing edit log: position=4, editlog=/opt/tmphadoop/dfs/name/current/edits
15/01/21 15:13:44 INFO namenode.FSEditLog: close success: truncate to 4, editlog=/opt/tmphadoop/dfs/name/current/edits
15/01/21 15:13:44 INFO common.Storage: Storage directory /opt/tmphadoop/dfs/name has been successfully formatted.
15/01/21 15:13:44 INFO namenode.NameNode: SHUTDOWN_MSG:
/************************************************************
SHUTDOWN_MSG: Shutting down NameNode at master/10.171.29.191
************************************************************/


2、启动hadoop【此步骤只需要在master上执行】

[jediael@master hadoop-1.2.1]$ bin/start-all.sh 

starting namenode, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-jediael-namenode-master.out
slave1: starting datanode, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-jediael-datanode-slave1.out
slave2: starting datanode, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-jediael-datanode-slave2.out
master: starting secondarynamenode, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-jediael-secondarynamenode-master.out
starting jobtracker, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-jediael-jobtracker-master.out
slave1: starting tasktracker, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-jediael-tasktracker-slave1.out
slave2: starting tasktracker, logging to /opt/jediael/hadoop-1.2.1/libexec/../logs/hadoop-jediael-tasktracker-slave2.out

3、登录页面验证
NameNode    http://ip:50070  
JobTracker     http://ip50030



4、查看各个主机的java进程
（1）master:
$ jps
17963 NameNode
18280 JobTracker
18446 Jps
18171 SecondaryNameNode
（2）slave1:
$ jps
16019 Jps
15858 DataNode
15954 TaskTracker
（3）slave2:
$ jps
15625 Jps
15465 DataNode
15561 TaskTracker

五、运行一个完整的mapreduce程序。

以下内容均只是master上执行
1、将wordcount.jar包复制至服务器上
程序见http://blog.csdn.net/jediael_lu/article/details/37596469

2、创建输入目录，并将相关文件复制至目录

[jediael@master166 ~]$ hadoop fs -mkdir /wcin
[jediael@master166 projects]$ hadoop fs -copyFromLocal /opt/jediael/hadoop-1.2.1/conf/hdfs-site.xml /wcin 

3、运行程序
[jediael@master166 projects]$ hadoop jar wordcount.jar org.jediael.hadoopdemo.wordcount.WordCount /wcin /wcout
14/08/31 20:04:26 WARN mapred.JobClient: Use GenericOptionsParser for parsing the arguments. Applications should implement Tool for the same.
14/08/31 20:04:26 INFO input.FileInputFormat: Total input paths to process : 1
14/08/31 20:04:26 INFO util.NativeCodeLoader: Loaded the native-hadoop library
14/08/31 20:04:26 WARN snappy.LoadSnappy: Snappy native library not loaded
14/08/31 20:04:26 INFO mapred.JobClient: Running job: job_201408311554_0003
14/08/31 20:04:27 INFO mapred.JobClient: map 0% reduce 0%
14/08/31 20:04:31 INFO mapred.JobClient: map 100% reduce 0%
14/08/31 20:04:40 INFO mapred.JobClient: map 100% reduce 100%
14/08/31 20:04:40 INFO mapred.JobClient: Job complete: job_201408311554_0003
14/08/31 20:04:40 INFO mapred.JobClient: Counters: 29
14/08/31 20:04:40 INFO mapred.JobClient: Job Counters
14/08/31 20:04:40 INFO mapred.JobClient: Launched reduce tasks=1
14/08/31 20:04:40 INFO mapred.JobClient: SLOTS_MILLIS_MAPS=4230
14/08/31 20:04:40 INFO mapred.JobClient: Total time spent by all reduces waiting after reserving slots (ms)=0
14/08/31 20:04:40 INFO mapred.JobClient: Total time spent by all maps waiting after reserving slots (ms)=0
14/08/31 20:04:40 INFO mapred.JobClient: Launched map tasks=1
14/08/31 20:04:40 INFO mapred.JobClient: Data-local map tasks=1
14/08/31 20:04:40 INFO mapred.JobClient: SLOTS_MILLIS_REDUCES=8531
14/08/31 20:04:40 INFO mapred.JobClient: File Output Format Counters
14/08/31 20:04:40 INFO mapred.JobClient: Bytes Written=284
14/08/31 20:04:40 INFO mapred.JobClient: FileSystemCounters
14/08/31 20:04:40 INFO mapred.JobClient: FILE_BYTES_READ=370
14/08/31 20:04:40 INFO mapred.JobClient: HDFS_BYTES_READ=357
14/08/31 20:04:40 INFO mapred.JobClient: FILE_BYTES_WRITTEN=104958
14/08/31 20:04:40 INFO mapred.JobClient: HDFS_BYTES_WRITTEN=284
14/08/31 20:04:40 INFO mapred.JobClient: File Input Format Counters
14/08/31 20:04:40 INFO mapred.JobClient: Bytes Read=252
14/08/31 20:04:40 INFO mapred.JobClient: Map-Reduce Framework
14/08/31 20:04:40 INFO mapred.JobClient: Map output materialized bytes=370
14/08/31 20:04:40 INFO mapred.JobClient: Map input records=11
14/08/31 20:04:40 INFO mapred.JobClient: Reduce shuffle bytes=370
14/08/31 20:04:40 INFO mapred.JobClient: Spilled Records=40
14/08/31 20:04:40 INFO mapred.JobClient: Map output bytes=324
14/08/31 20:04:40 INFO mapred.JobClient: Total committed heap usage (bytes)=238026752
14/08/31 20:04:40 INFO mapred.JobClient: CPU time spent (ms)=1130
14/08/31 20:04:40 INFO mapred.JobClient: Combine input records=0
14/08/31 20:04:40 INFO mapred.JobClient: SPLIT_RAW_BYTES=105
14/08/31 20:04:40 INFO mapred.JobClient: Reduce input records=20
14/08/31 20:04:40 INFO mapred.JobClient: Reduce input groups=20
14/08/31 20:04:40 INFO mapred.JobClient: Combine output records=0
14/08/31 20:04:40 INFO mapred.JobClient: Physical memory (bytes) snapshot=289288192
14/08/31 20:04:40 INFO mapred.JobClient: Reduce output records=20
14/08/31 20:04:40 INFO mapred.JobClient: Virtual memory (bytes) snapshot=1533636608
14/08/31 20:04:40 INFO mapred.JobClient: Map output records=20

4、查看结果

[jediael@master166 projects]$ hadoop fs -cat /wcout/* 

--> 1
<!-- 1
</configuration> 1
</property> 1
<?xml 1
<?xml-stylesheet 1
<configuration> 1
<name>dfs.replication</name> 1
<property> 1
<value>2</value> 1
Put 1
file. 1
href="configuration.xsl"?> 1
in 1
overrides 1
property 1
site-specific 1
this 1
type="text/xsl" 1
version="1.0"?> 1
cat: File does not exist: /wcout/_logs