hadoop集群搭建

最新推荐文章于 2025-08-11 17:27:24 发布

原创最新推荐文章于 2025-08-11 17:27:24 发布 · 2.4k 阅读

0 ·

CC 4.0 BY-SA版权

本文为互联网老辛原创文章，未经博主允许不得转载

环境：
在这里插入图片描述

配好IP，关闭iptables，关闭selinux,配置hosts
配置hosts:

192.168.1.18 xinsz08-1
192.168.1.20 xinsz08-64
192.168.1.11 xinsz08-62

三台服务器上都要创建普通用户，hadoop，配置密码： 123456

[root@xinsz08-64 ~]# useradd -u 8000 hadoop ; echo 123456 | passwd --stdin hadoop
更改用户 hadoop 的密码 。
passwd： 所有的身份验证令牌已经成功更新。

设置namenode能够无密钥登录另外两台服务器

[root@xinsz08-1 ~]# ssh-keygen

Generating public/private rsa key pair.
Enter file in which to save the key (/root/.ssh/id_rsa): 
Enter passphrase (empty for no passphrase): 
Enter same passphrase again: 
Your identification has been saved in /root/.ssh/id_rsa.
Your public key has been saved in /root/.ssh/id_rsa.pub.
The key fingerprint is:
14:35:ad:7d:68:de:9a:4d:e0:83:06:e4:5a:d4:c5:42 root@xinsz08-1
The key's randomart image is:
+--[ RSA 2048]----+
|        .+E+.    |
|        o.o.o    |
|       +.  + .   |
|       .+ . = .  |
|       oS. = +   |
|      .   o + o  |
|         .   *   |
|            o .  |
|                 |
+-----------------+

[root@xinsz08-1 ~]# ssh-copy-id  root@192.168.1.11
The authenticity of host '192.168.1.11 (192.168.1.11)' can't be established.
RSA key fingerprint is 3f:70:d9:6f:c8:54:70:6f:14:0f:73:10:44:fc:51:7d.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added '192.168.1.11' (RSA) to the list of known hosts.
root@192.168.1.11's password: 

[root@xinsz08-1 ~]# ssh-copy-id  root@192.168.1.20

测试（是否能登录成功）
[root@xinsz08-1 ~]# ssh 192.168.1.20
Last login: Tue Mar  3 13:36:50 2020 from 192.168.1.4

安装jdk

[root@xinsz08-1 ~]# rpm -ivh jdk-8u131-linux-x64_.rpm


//查看java的安装目录
195  rpm -pql /root/jdk-8u131-linux-x64_.rpm

修改/etc/profile

export JAVA_HOME=/usr/java/jdk1.8.0_131/
export JAVA_BIN=/usr/java/jdk1.8.0_131/bin/
export PATH=${JAVA_HOME}/bin:$PATH
export CLASSPATH=.:${JAVA_HOME}/lib/dt.jar:${JAVA_HOME}/lib/tools.jar

//j加载：

[root@xinsz08-1 ~]# source /etc/profile
[root@xinsz08-1 ~]# java -version
java version "1.8.0_131"
Java(TM) SE Runtime Environment (build 1.8.0_131-b11)
Java HotSpot(TM) 64-Bit Server VM (build 25.131-b11, mixed mode)
[root@xinsz08-1 ~]#

查看java版本

[root@xinsz08-1 native]# java -version
java version "1.8.0_131"
Java(TM) SE Runtime Environment (build 1.8.0_131-b11)
Java HotSpot(TM) 64-Bit Server VM (build 25.131-b11, mixed mode)
[root@xinsz08-1 native]#

问题： source 在数据库里还可以用来做什么？
答：导入

两个datanode节点也需要安装java，可以直接从主节点拷贝过去

  203  scp jdk-8u131-linux-x64_.rpm root@192.168.1.11:/root
  204  scp jdk-8u131-linux-x64_.rpm root@192.168.1.20:/root
  205  scp /etc/profile  root@192.168.1.20:/etc/profile
  206  scp /etc/profile  root@192.168.1.11:/etc/profile

在另外两个节点安装java（一台192.168.1.11.另一台192.168.1.20）

[root@xinsz08-64 ~]# rpm -ivh jdk-8u131-linux-x64_.rpm 

```bash
[root@xinsz08-62 ~]# source /etc/profile
[root@xinsz08-62 ~]# java -version
java version "1.8.0_131"
Java(TM) SE Runtime Environment (build 1.8.0_131-b11)
Java HotSpot(TM) 64-Bit Server VM (build 25.131-b11, mixed mode)

6. 创建hadoop用户（三台机器都需要创建hadoop）

```bash
[root@xinsz08-64 ~]# useradd -u 8000 hadoop ; echo 123456 | passwd --stdin hadoop
更改用户 hadoop 的密码 。
passwd： 所有的身份验证令牌已经成功更新。

安装namenode

Hadoop 安装目录：/home/hadoop/hadoop-3.13 使用 root 帐号将 hadoop-3.1.3.tar.gz 上传到服务器，并且放到/home/hadoop下！

1）创建dfs和tmp
rz

[root@xinsz08-1 hadoop]# su - hadoop
[hadoop@xinsz08-1 ~]$ whoami
hadoop

$ mkdir -p /home/hadoop/dfs/name /home/hadoop/dfs/data /home/hadoop/tmp

2）解压

[hadoop@xinsz08-1 ~]$ tar xvf hadoop-3.1.3.tar.gz

[hadoop@xinsz08-1 ~]$ cd hadoop-3.1.3
[hadoop@xinsz08-1 hadoop-3.1.3]$ ll
总用量 200
drwxr-xr-x. 2 hadoop hadoop   4096 9月  12 12:46 bin
drwxr-xr-x. 3 hadoop hadoop   4096 9月  12 10:51 etc
drwxr-xr-x. 2 hadoop hadoop   4096 9月  12 12:46 include
drwxr-xr-x. 3 hadoop hadoop   4096 9月  12 12:46 lib
drwxr-xr-x. 4 hadoop hadoop   4096 9月  12 12:46 libexec
-rw-rw-r--. 1 hadoop hadoop 147145 9月   4 17:31 LICENSE.txt
-rw-rw-r--. 1 hadoop hadoop  21867 9月   4 17:31 NOTICE.txt
-rw-rw-r--. 1 hadoop hadoop   1366 9月   4 17:31 README.txt
drwxr-xr-x. 3 hadoop hadoop   4096 9月  12 10:51 sbin
drwxr-xr-x. 4 hadoop hadoop   4096 9月  12 13:08 share

[hadoop@xinsz08-1 hadoop]$ pwd
/home/hadoop/hadoop-3.1.3/etc/hadoop
[hadoop@xinsz08-1 hadoop]$ ls

hadoop-env.sh  // java的环境变量
yarn-env.sh  //制定yarn框架的Java运行环境
slaves//指定datanode数据存储服务器
core-site.xml     // hadoop-web界面路径
hdfs-site.xml   //文件系统的配置文件
mapred-site.xml// mapreducer 任务配置文件
yarn-site.xml // yarn框架配置，主要一些任务的启动位置
                                    
[hadoop@xinsz08-1 hadoop]$

3）修改配置文件


[hadoop@xinsz08-1 hadoop]$ vim hadoop-env.sh 
 38 exprot JAVA_HOME=/usr/java/jdk1.8.0_13

[hadoop@xinsz08-1 hadoop]$ vim yarn-env.sh
 27 JAVA_HOME=/usr/java/jdk1.8.0_131

[hadoop@xinsz-01 hadoop]$ vim slaves
xinsz08-64
xinsz08-62

[hadoop@xinsz08-1 hadoop]$ vim core-site.xml

备注：这个是hadoop的核心配置，这里需要配置两属性， fs.default.name 配置hadoop的HDFS系统命令，位置为主机的9000端口， hadoop.tmp.dir 配置haddop的tmp目录的根位置。

<configuration>
<property>
        <name>fs.defaultFS</name>
        <value>hdfs://xinsz08-1:9000</value>
</property>

<property>
        <name>io.file.buffer.size</name>
        <value>131072</value>
</property>


<property>
        <name>hadoop.tmp.dir</name>
        <value>file:/home/hadoop/tmp</value>
        <description>Abase for other tmporary directries.</description>
</property>

[hadoop@xinsz08-1 hadoop]$ vim hdfs-site.xml

备注： HDFS主要的配置文件， dfs.http.address配置了hdfs的http的访问位置；
dfs.replication 配置文件的副本，一般不大于从机个数

<configuration>
<property>
	<name>dfs.namenode.secondary.http-address</name>
	<value>xinsz08-1:9000</value>
</property>

<property>
	<name>dfs.namenode.name.dir</name>
	<value>file:/home/hadoop/dfs/name</value>
</property>

<property>
	<name>dfs.datanode.data.dir</name>
	<value>file:/home/hadoop/dfs/data</value>
</property>

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

<property>
	<name>dfs.webhdfs.enabled</name>
	<value>true</value>
</property>
</configuration>

[hadoop@xinsz08-1 hadoop]$ vim mapred-site.xml

备注：这个是mapreduce任务配置文件，mapreduce.framework.name 属性下配置yarn,
mapred.map.tasks和mapred.reduce.tasks 分别为map和reduce 的任务数。同时指定hadoop历史服务器hsitoryserver

我们可以通过historyserver查看mapreduce的作业记录，比如用了多少个map,用了多少个reduce，作业启动时间，作业完成时间。默认清空下，hadoop历史服务器是没有启动的，我们需要通过命令来启动

[hadoop@xinsz08-1 hadoop]$ /home/hadoop/hadoop-3.1.3/sbin/mr-jobhistory-daemon.sh  start historyserver
/home/hadoop/hadoop-3.1.3/etc/hadoop/hadoop-env.sh: line 38: exprot: command not found
WARNING: Use of this script to start the MR JobHistory daemon is deprecated.
WARNING: Attempting to execute replacement "mapred --daemon start" instead.
WARNING: /home/hadoop/hadoop-3.1.3/logs does not exist. Creating.
[hadoop@xinsz08-1 hadoop]$

<configuration>
<property>
	<name>mapreduce.framework.name</name>
	<value>yarn</value>
</property>

<property>
	<name>mapreduce.jobhistory.address</name>
	<value>xinsz08-1:10020</value>
</property>

<property>
	<name>mapreduce.jobhistory.webapp.address</name>
	<value>xinsz08-1:19888</value>
</property>

</configuration>

[hadoop@xinsz08-1 hadoop]$ vim yarn-site.xml
备注： yarn框架的配置，主要是一些任务的启动位置

<configuration>

<!-- Site specific YARN configuration properties -->
<proetry>
	<name>yarn.nodemanager.aux-service</name>
	<value>mapreduce_shuffle</value>
</proetry>

<proetry>
	<name>yarn.nodemanager.uax-service.mapreduce.shuffle.class</name>
	<value>org.apache.hadoop.mapreduced.ShuffleHandle</value>
</proetry>
<proetry>
	<name>yarn.resoucemanager.address</name>
	<value>xinsz08-1:8032</value>
</proetry>
<proetry>
	<name>yarn.resourcemanager.shceduler.address</name>
	<value>xinsz08-1:8030</value>
</proetry>
<proetry>
	<name>yarn.resourcemanager.resource-tracker.address</name>
	<value>xinsz08-1:8031</value>
</proetry>

<proetry>
	<name>yarn.resourcemanager.admin.address</name>
	<value>xinsz08-1:8033</value>
</proetry>

<proetry>
	<name>yarn.resourcemanager.webapp.address</name>
	<value>xinsz08-1:8088</value>
</proetry>
</configuration>

datanode配置文件生成
[hadoop@xinsz-01 hadoop]$ scp -r /home/hadoop/hadoop-3.13 hadoop@xinsz08-64:~/
[hadoop@xinsz-01 hadoop]$ scp -r /home/hadoop/hadoop-3.13 hadoop@xinsz08-62:~/

5） namenode格式化数据：
一般第一次的时候需要初始化，之后就不需要了

2020-03-04 16:05:17,247 INFO namenode.FSImageFormatProtobuf: Image file /home/hadoop/dfs/name/current/fsimage.ckpt_0000000000000000000 of size 393 bytes saved in 0 seconds .
2020-03-04 16:05:17,268 INFO namenode.NNStorageRetentionManager: Going to retain 1 images with txid >= 0
2020-03-04 16:05:17,277 INFO namenode.FSImage: FSImageSaver clean checkpoint: txid = 0 when meet shutdown.
2020-03-04 16:05:17,278 INFO namenode.NameNode: SHUTDOWN_MSG: 
/************************************************************
SHUTDOWN_MSG: Shutting down NameNode at xinsz08-1/192.168.1.18
************************************************************/

6）查看是否生成相应的内容

[hadoop@xinsz08-1 ~]$ cd /home/hadoop/dfs/
[hadoop@xinsz08-1 dfs]$ ls
data  name
[hadoop@xinsz08-1 dfs]$ tree 
.
├── data
└── name
    └── current
        ├── fsimage_0000000000000000000
        ├── fsimage_0000000000000000000.md5
        ├── seen_txid
        └── VERSION

3 directories, 4 files

7）配置免密要登录

[hadoop@xinsz08-1 dfs]$ ssh-keygen 
Generating public/private rsa key pair.
Enter file in which to save the key (/home/hadoop/.ssh/id_rsa): 
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:
cb:ff:ad:50:92:fb:38:cb:1d:93:ed:7e:3f:54:d7:cd hadoop@xinsz08-1
The key's randomart image is:
+--[ RSA 2048]----+
|                 |
|                 |
|               .o|
|           .    E|
|        S o .   o|
|       . . + o . |
|        o o + o  |
|         o.= = ..|
|          =+=o+.+|
+-----------------+
[hadoop@xinsz08-1 dfs]$

[hadoop@xinsz-01 hadoop]$ ssh-keygen
[hadoop@xinsz-01 hadoop]$ ssh-copy-id xinsz08-64
[hadoop@xinsz-01 hadoop]$ ssh-copy-id xinsz08-1 //对自己也做一次
[hadoop@xinsz-01 hadoop]$ ssh-copy-id wg61.cn

备注：方便后期复制文件或者启动服务。因为namenode启动时候，会链接到datanode上启动对应的服务，

启动hdfs
[hadoop@xinsz08-1 hadoop]$ pwd
/home/hadoop/hadoop-3.1.3/etc/hadoop

报错：2020-03-04 16:16:45,394 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform… using builtin-java classes where applicable

解答： http://dl.bintray.com/sequenceiq/sequenceiq-bin/ 下载对应版本
解压，覆盖hadoop下/lib/native/
上传之后解压：

[hadoop@xinsz08-1 native]$ tar xf hadoop-native-64.tar 
[hadoop@xinsz08-1 native]$ ls
examples              libhadoop.so.1.0.0  libnativetask.a
hadoop-native-64.tar  libhadooputils.a    libnativetask.so
libhadoop.a           libhdfs.a           libnativetask.so.1.0.0
libhadooppipes.a      libhdfs.so
libhadoop.so          libhdfs.so.0.0.0

覆盖完之后重启
[hadoop@xinsz08-1 hadoop]$ …/…/sbin/stop-dfs.sh
关闭之后在启动

[hadoop@xinsz08-1 hadoop]$ ../../sbin/start-dfs.sh
Starting namenodes on [xinsz08-1]
Starting datanodes
Starting secondary namenodes [xinsz08-1]

启动yarn
也就是说我们要启动分布式计算

[hadoop@xinsz08-1 hadoop]$ pwd
/home/hadoop/hadoop-3.1.3/etc/hadoop

[hadoop@xinsz08-1 hadoop]$ ../../sbin/start-yarn.sh

除了使用上面的脚本之外，我们还可以使用：

[hadoop@xinsz08-1 hadoop]$ ../../sbin/start-all.sh

启动jobhistory
…/…/sbin/mr-jobhistory-daemon.sh start historyserver
启动存储服务和资源管理主服务

[hadoop@xinsz08-1 hadoop-3.1.3]$ sbin/hadoop-daemon.sh  start datanode
WARNING: Use of this script to start HDFS daemons is deprecated.
WARNING: Attempting to execute replacement "hdfs --daemon start" instead.
[hadoop@xinsz08-1 hadoop-3.1.3]$ sbin/yarn-daemon.sh start nodemanager
WARNING: Use of this script to start YARN daemons is deprecated.
WARNING: Attempting to execute replacement "yarn --daemon start" instead.
[hadoop@xinsz08-1 hadoop-3.1.3]$