hadoop集群搭建(hadoop)_现有20台服务器,如何做集群-优快云博客

本文链接：https://blog.youkuaiyun.com/lengzijian/article/details/7042181

本文详细介绍了Hadoop集群的环境配置、SSH免密登录设置、版本选择、配置文件编辑、启动流程以及验证Hadoop文件系统搭建成功的步骤。通过本地浏览器查看集群状态，确保系统正常运行。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

首先说一下配置环境：三台电脑
192.168.30.149  hadoop149 namenode和jobtracker   ###因为149机器稍微好一点
192.168.30.150  hadoop150 datanode和TaskTracker
192.168.30.148  hadoop150 datanode和TaskTracker

配置ssh无需密码登陆：

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

我的master在149可以吧149的.pub文件拷贝到150和148上 然后执行cat~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys

如果存在问题有可能是文件权限问题！

我用的hadoop版本是hadoop-0.20.2 下载地址：

google吧 过两天弄个网盘都放在上面再写到这里。
下载后：编辑几个文件：
在/root/hadoop-0.20.2/conf中(这里注意的是几台电脑的hadoop文件路径必须相同)：加入如下一句话
   [root@localhostconf]# vim hadoop-env.sh       
 exportJAVA_HOME=/usr/java/jdk1.7.0_01     ###设置变量

[root@localhostconf]# vim core-site.xml

<?xmlversion="1.0"?>
<?xml-stylesheettype="text/xsl" href="configuration.xsl"?>
        
<!-- Put site-specificproperty overrides in this file. -->
        
<configuration>
<property>
<name>fs.default.name</name>
<value>hdfs://192.168.30.149:9000</value> ###具体的意义之后会讲解
</property>
</configuration>

[root@localhostconf]# vim mapred-site.xml

<?xmlversion="1.0"?>
<?xml-stylesheettype="text/xsl" href="configuration.xsl"?>
 
<!-- Putsite-specific property overrides in this file. -->
 
<configuration>
<property>
<name>mapred.job.tracker</name>
<value>hdfs://192.168.30.149:9004</value>
</property>
</configuration>

[root@localhostconf]# vim hdfs-site.xml

<?xmlversion="1.0"?>
<?xml-stylesheettype="text/xsl" href="configuration.xsl"?>
 
<!-- Putsite-specific property overrides in this file. -->
 
<configuration>
<property>
<name>dfs.replication</name>
<value>2</value>
</property>
</configuration>

[root@localhostconf]# vim masters

hadoop149

[root@localhostconf]# vim slaves

hadoop150
hadoop148

一共编辑了5个文件,具体意义代表什么，之后会讲到

这里注意要被指/etc/hosts文件,如下(192.168.30.149)：

[root@localhostconf]# vim /etc/hosts

# Do not removethe following line, or various programs
# that requirenetwork functionality will fail.
127.0.0.1               localhost.localdomain localhost
::1             localhost6.localdomain6 localhost6
        
192.168.30.149hadoop149
192.168.30.150hadoop150
192.168.30.148hadoop148

4.启动hadoop：

这里用简单的命令进行启动，

A．格式化文件系统：

 #bin/hadoop namenode –format

B．启动hadoop

  #bin/start-all.sh

C．利用hadoop自带的例子测试hadoop是否启动成功

#bin/hadoop fs -mkdir input     ###在文件系统中创建input文件夹
#bin/hadoopfs -put README.txt input    ###把本地readme.txt上传到input中
#bin/hadoop fs –lsr            ###查看本件系统所有文件
    存在文件并且大小不为0则hadoop文件系统搭建成功。
#bin/hadoopjar hadoop-0.20.2-examples.jar wordcount input/README.txt output
                                                                                    ###将输出结果输出到output中
#bin/hadoop jar hadoop-0.20.2-examples.jar wordcount input/1.txt output

11/12/02 17:47:14 INFOinput.FileInputFormat: Total input paths to process : 1

11/12/02 17:47:14 INFO mapred.JobClient:Running job: job_201112021743_0001

11/12/02 17:47:15 INFOmapred.JobClient: map 0% reduce 0%

11/12/02 17:47:22 INFOmapred.JobClient: map 100% reduce 0%

11/12/02 17:47:34 INFOmapred.JobClient: map 100% reduce 100%

11/12/02 17:47:36 INFO mapred.JobClient:Job complete: job_201112021743_0001

11/12/02 17:47:36 INFO mapred.JobClient:Counters: 17

11/12/02 17:47:36 INFOmapred.JobClient: Job Counters

11/12/02 17:47:36 INFOmapred.JobClient: Launched reducetasks=1

11/12/02 17:47:36 INFOmapred.JobClient: Launched maptasks=1

11/12/02 17:47:36 INFOmapred.JobClient: Data-local maptasks=1

11/12/02 17:47:36 INFOmapred.JobClient: FileSystemCounters

11/12/02 17:47:36 INFOmapred.JobClient: FILE_BYTES_READ=32523

11/12/02 17:47:36 INFOmapred.JobClient: HDFS_BYTES_READ=44253

11/12/02 17:47:36 INFOmapred.JobClient: FILE_BYTES_WRITTEN=65078

11/12/02 17:47:36 INFOmapred.JobClient: HDFS_BYTES_WRITTEN=23148

11/12/02 17:47:36 INFOmapred.JobClient: Map-Reduce Framework

11/12/02 17:47:36 INFOmapred.JobClient: Reduce inputgroups=2367

11/12/02 17:47:36 INFOmapred.JobClient: Combine outputrecords=2367

11/12/02 17:47:36 INFOmapred.JobClient: Map inputrecords=734

11/12/02 17:47:36 INFOmapred.JobClient: Reduce shufflebytes=32523

11/12/02 17:47:36 INFOmapred.JobClient: Reduce outputrecords=2367

11/12/02 17:47:36 INFO mapred.JobClient: Spilled Records=4734

11/12/02 17:47:36 INFOmapred.JobClient: Map outputbytes=73334

11/12/02 17:47:36 INFOmapred.JobClient: Combine inputrecords=7508

11/12/02 17:47:36 INFOmapred.JobClient: Map outputrecords=7508

11/12/02 17:47:36 INFOmapred.JobClient: Reduce inputrecords=2367

也可以通过本地浏览器进行查看状态：50070和50030端口（注意配置本地C:\Windows\System32\drivers\etc\hosts文件）

192.168.30.150      hadoop150
192.168.30.149      hadoop149
192.168.30.148      hadoop148