hadoop 多机全分布式安装步骤（虚拟机1master+2slave）

文章目录

1. 虚拟机安装Centos7
2. 配置静态IP
3. 更改主机名
4. 编辑域名映射
5. 安装配置Java
6. 配置SSH免密登录
7 .安装Hadoop
8. 关闭防火墙
9. 格式化文件系统
10. 启动验证
11. 第一个MapReduce程序: WordCount
12. 关闭Hadoop

参考书：《Hadoop大数据原理与应用》

1. 虚拟机安装Centos7

安装3台虚拟机，centos7，一个master，两个slave，安装时可以改hostname, 记得设置密码
安装的是4.7Gb的包，选择的 service with GUI
选则 NAT 网络链接
ip route show 查看路由器网关ip
ip addr 查找本机ip（下面用的着这两个ip）

2. 配置静态IP

vim /etc/sysconfig/network-scripts/ifcfg-ens33

TYPE=Ethernet
PROXY_METHOD=none
BROWSER_ONLY=no
BOOTPROTO=static # 改静态
DEFROUTE=yes
IPV4_FAILURE_FATAL=no
IPV6INIT=yes
IPV6_AUTOCONF=yes
IPV6_DEFROUTE=yes
IPV6_FAILURE_FATAL=no
IPV6_ADDR_GEN_MODE=stable-privacy
NAME=ens33
UUID=caf90547-4b5a-46b3-ab7c-2c8fb1f5e4d7
DEVICE=ens33
ONBOOT=yes # 改yesIPADDR=192.168.253.130 # ip
NETMASK=255.255.255.0
GATEWAY=192.168.253.2 # 网关
DNS1=192.168.253.2 # 跟网关一样即可

保存权限不足，输入w !sudo tee %

重启网络

systemctl restart network.service

2: ens33: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000link/ether 00:0c:29:77:14:71 brd ff:ff:ff:ff:ff:ffinet 192.168.253.130/24 brd 192.168.253.255 scope global noprefixroute ens33valid_lft forever preferred_lft foreverinet6 fe80::ce06:b26c:ff2d:5288/64 scope link noprefixroute valid_lft forever preferred_lft forever

同理，另外两台 ip 为：192.168.253.128, 192.168.253.129（个人根据自己的情况来）

3. 更改主机名

安装的时候就改了，此处可跳过
切换 root 用户，sudo su
vi /etc/hostname, 分别替换内容为 master，slave1， slave2
reboot重启，hostname 查看是否更改

4. 编辑域名映射

为了便捷访问，三台机器都做以下修改，sudo su
在 /etc/hosts追加以下内容，重启

192.168.253.130 master
192.168.253.128 slave1
192.168.253.129 slave2
192.168.31.237 michael

检查各台机器是否能ping通

ping master
ping slave1
ping slave2

5. 安装配置Java

卸载
查看 java -version
卸载自带的 Oracle OpenJDK，使用Oracle JDK

rpm -qa | grep jdk

yum -y remove  java-1.8.0*
yum -y remove  java-1.7.0*

下载 jdk，位数根据下图来
我从宿主机直接考过来安装包

scp root@michael:/home/dnn/jdk-8u281-linux-x64.tar.gz /home/dnn/jdk-8u281-linux-x64.tar.gz

参考 JDK 安装
装到/opt/jdk1.8.0_281/

6. 配置SSH免密登录

查询 rpm -qa | grep ssh
没有的话安装

yum -y install openssh
yum -y install openssh-server
yum -y install openssh-clients

vim /etc/ssh/sshd_config
第43行取消注释，并加一行，3台机器都做

RSAAuthentication yes
PubkeyAuthentication yes

systemctl restart sshd.service，重启服务
切换普通用户 ctrl+d，回到home cd ~
ssh-keygen,一直回车
cat ~/.ssh/id_rsa.pub >> ~/.ssh/authorized_keys
chmod 0600 ~/.ssh/authorized_keys
将master的公钥复制给slave1，slave2，免密访问从节点

在master里键入以下命令

ssh-copy-id -i ~/.ssh/id_rsa.pub dnn@slave1
ssh slave1
ssh-copy-id -i ~/.ssh/id_rsa.pub dnn@slave2
ssh slave2
ssh master

遇到提示输入 yes，敲密码

还可以在另外两台里，同样的步骤操作一遍

7 .安装Hadoop

伪分布式可以参考：hadoop 单机伪分布式安装步骤

下载或拷贝安装包到3台机器

scp dnn@michael:/home/dnn/hadoop-3.3.0.tar.gz /home/dnn/hadoop-3.3.0.tar.gz

到文件目录下，解压 tar -zxvf hadoop-3.3.0.tar.gz
移动到你要放的目录 sudo mv hadoop-3.3.0 /opt/hadoop-3.3.0
赋权限给普通用户dnn，chown -R dnn /opt/hadoop-3.3.0

在主节点上操作：

切换 root 用户，新建文件 vim /etc/profile.d/hadoop.sh
添加内容

export HADOOP_HOME=/opt/hadoop-3.3.0
export PATH=$HADOOP_HOME/bin:$HADOOP_HOME/sbin:$PATH

source /etc/profile.d/hadoop.sh

切换普通用户，上面已赋权限，vim /opt/hadoop-3.3.0/etc/hadoop/hadoop-env.sh
54行改为 export JAVA_HOME=/opt/jdk1.8.0_281/
55行添加 export HADOOP_SSH_OPTS='-o StrictHostKeyChecking=no'
199行修改 export HADOOP_PID_DIR=${HADOOP_HOME}/pids
vim /opt/hadoop-3.3.0/etc/hadoop/mapred-env.sh
添加 export JAVA_HOME=/opt/jdk1.8.0_281/，export HADOOP_MAPRED_PID_DIR=${HADOOP_HOME}/pids
vim /opt/hadoop-3.3.0/etc/hadoop/yarn-env.sh
添加

export JAVA_HOME=/opt/jdk1.8.0_281/
export YARN_PID_DIR=${HADOOP_HOME}/pids

vim /opt/hadoop-3.3.0/etc/hadoop/core-site.xml

<configuration><property><name>fs.defaultFS</name><value>hdfs://192.168.253.130:9000</value></property><property><name>hadoop.tmp.dir</name><value>/opt/hadoop-3.3.0/hdfsdata</value></property><property><name>io.file.buffer.size</name><value>131072</value></property>
<configuration>

vim /opt/hadoop-3.3.0/etc/hadoop/mapred-site.xml

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

vim /opt/hadoop-3.3.0/etc/hadoop/yarn-site.xml

<configuration><property><name>yarn.resourcemanager.hostname</name><value>master</value></property><property><name>yarn.nodemanager.aux-services</name><value>mapreduce_shuffle</value></property>
</configuration>

在 /opt/hadoop-3.3.0/etc/hadoop/ 下，vim workers
删除 localhost，加入

slave1
slave2

同步配置文件到2台slave上

在root下

scp /etc/profile.d/hadoop.sh root@slave1:/etc/profile.d/
scp /etc/profile.d/hadoop.sh root@slave2:/etc/profile.d/

在普通用户下

scp -r /opt/hadoop-3.3.0/etc/hadoop/* dnn@slave1:/opt/hadoop-3.3.0/etc/hadoop/
scp -r /opt/hadoop-3.3.0/etc/hadoop/* dnn@slave2:/opt/hadoop-3.3.0/etc/hadoop/

8. 关闭防火墙

在root下，systemctl disable firewalld.service
重启，再看下状态 systemctl status firewalld.service
显示 inactive(dead), 3台机器都做

9. 格式化文件系统

只在 master 上用普通用户操作：

hdfs namenode -format

10. 启动验证

在 master 上执行3条命令

start-dfs.sh
start-yarn.sh
mr-jobhistory-daemon.sh start historyserver
# 第三条可以用下面的命令，上面的显示过期了，以后弃用
mapred --daemon start historyserver

输入 jps 命令，可以看见进程启动了

11. 第一个MapReduce程序: WordCount

[dnn@master ~]$ hdfs dfs -mkdir /InputDataTest
[dnn@master ~]$ hdfs dfs -ls /
Found 2 items
drwxr-xr-x   - dnn supergroup          0 2021-03-12 06:49 /InputDataTest
drwxrwx---   - dnn supergroup          0 2021-03-12 06:19 /tmp

上传文件到 InputDataTest 文件夹

[dnn@master ~]$ hdfs dfs -put /opt/hadoop-3.3.0/etc/hadoop/hadoop-env.sh /InputDataTest
[dnn@master ~]$ hdfs dfs -put /opt/hadoop-3.3.0/etc/hadoop/mapred-env.sh /InputDataTest
[dnn@master ~]$ hdfs dfs -put /opt/hadoop-3.3.0/etc/hadoop/yarn-env.sh /InputDataTest
[dnn@master ~]$ hdfs dfs -ls /InputDataTest
Found 3 items
-rw-r--r--   3 dnn supergroup      17017 2021-03-12 06:52 /InputDataTest/hadoop-env.sh
-rw-r--r--   3 dnn supergroup       1850 2021-03-12 06:53 /InputDataTest/mapred-env.sh
-rw-r--r--   3 dnn supergroup       6406 2021-03-12 06:53 /InputDataTest/yarn-env.sh

hadoop jar /opt/hadoop-3.3.0/share/hadoop/mapreduce/hadoop-mapreduce-examples-3.3.0.jar wordcount /InputDataTest /OutputDataTest
报错： org.apache.hadoop.mapreduce.v2.app.MRAppMaster

vim /opt/hadoop-3.3.0/etc/hadoop/yarn-site.xml

<configuration><!-- Site specific YARN configuration properties --><property><name>yarn.resourcemanager.hostname</name><value>master</value></property><property><name>yarn.nodemanager.aux-services</name><value>mapreduce_shuffle</value></property><property><name>yarn.application.classpath</name><value>/opt/hadoop-3.3.0/etc/hadoop:/opt/hadoop-3.3.0/share/hadoop/common/lib/*:/opt/hadoop-3.3.0/share/hadoop/common/*:/opt/hadoop-3.3.0/share/hadoop/hdfs:/opt/hadoop-3.3.0/share/hadoop/hdfs/lib/*:/opt/hadoop-3.3.0/share/hadoop/hdfs/*:/opt/hadoop-3.3.0/share/hadoop/mapreduce/*:/opt/hadoop-3.3.0/share/hadoop/yarn:/opt/hadoop-3.3.0/share/hadoop/yarn/lib/*:/opt/hadoop-3.3.0/share/hadoop/yarn/*
</value></property>
</configuration>

重启集群（关闭3条命令，见下面第12节，启动3条命令），再次运行 wordcount 程序

[dnn@master ~]$ hadoop jar /opt/hadoop-3.3.0/share/hadoop/mapreduce/hadoop-mapreduce-examples-3.3.0.jar wordcount /InputDataTest /OutputDataTest
2021-03-12 07:11:51,635 INFO client.DefaultNoHARMFailoverProxyProvider: Connecting to ResourceManager at master/192.168.253.130:8032
2021-03-12 07:11:52,408 INFO mapreduce.JobResourceUploader: Disabling Erasure Coding for path: /tmp/hadoop-yarn/staging/dnn/.staging/job_1615504213995_0001
2021-03-12 07:11:53,547 INFO input.FileInputFormat: Total input files to process : 3
2021-03-12 07:11:54,066 INFO mapreduce.JobSubmitter: number of splits:3
2021-03-12 07:11:54,271 INFO mapreduce.JobSubmitter: Submitting tokens for job: job_1615504213995_0001
2021-03-12 07:11:54,271 INFO mapreduce.JobSubmitter: Executing with tokens: []
2021-03-12 07:11:54,624 INFO conf.Configuration: resource-types.xml not found
2021-03-12 07:11:54,624 INFO resource.ResourceUtils: Unable to find 'resource-types.xml'.
2021-03-12 07:11:55,117 INFO impl.YarnClientImpl: Submitted application application_1615504213995_0001
2021-03-12 07:11:55,164 INFO mapreduce.Job: The url to track the job: http://master:8088/proxy/application_1615504213995_0001/
2021-03-12 07:11:55,164 INFO mapreduce.Job: Running job: job_1615504213995_0001
2021-03-12 07:12:05,308 INFO mapreduce.Job: Job job_1615504213995_0001 running in uber mode : false
2021-03-12 07:12:05,319 INFO mapreduce.Job:  map 0% reduce 0%
2021-03-12 07:12:21,455 INFO mapreduce.Job:  map 33% reduce 0%
2021-03-12 07:12:22,460 INFO mapreduce.Job:  map 100% reduce 0%
2021-03-12 07:12:29,514 INFO mapreduce.Job:  map 100% reduce 100%
2021-03-12 07:12:29,526 INFO mapreduce.Job: Job job_1615504213995_0001 completed successfully
2021-03-12 07:12:29,652 INFO mapreduce.Job: Counters: 54File System CountersFILE: Number of bytes read=20470FILE: Number of bytes written=1097885FILE: Number of read operations=0FILE: Number of large read operations=0FILE: Number of write operations=0HDFS: Number of bytes read=25631HDFS: Number of bytes written=12134HDFS: Number of read operations=14HDFS: Number of large read operations=0HDFS: Number of write operations=2HDFS: Number of bytes read erasure-coded=0Job Counters Launched map tasks=3Launched reduce tasks=1Data-local map tasks=3Total time spent by all maps in occupied slots (ms)=42362Total time spent by all reduces in occupied slots (ms)=4808Total time spent by all map tasks (ms)=42362Total time spent by all reduce tasks (ms)=4808Total vcore-milliseconds taken by all map tasks=42362Total vcore-milliseconds taken by all reduce tasks=4808Total megabyte-milliseconds taken by all map tasks=43378688Total megabyte-milliseconds taken by all reduce tasks=4923392Map-Reduce FrameworkMap input records=667Map output records=3682Map output bytes=39850Map output materialized bytes=20482Input split bytes=358Combine input records=3682Combine output records=1261Reduce input groups=912Reduce shuffle bytes=20482Reduce input records=1261Reduce output records=912Spilled Records=2522Shuffled Maps =3Failed Shuffles=0Merged Map outputs=3GC time elapsed (ms)=800CPU time spent (ms)=2970Physical memory (bytes) snapshot=615825408Virtual memory (bytes) snapshot=10951270400Total committed heap usage (bytes)=385785856Peak Map Physical memory (bytes)=168960000Peak Map Virtual memory (bytes)=2738552832Peak Reduce Physical memory (bytes)=110534656Peak Reduce Virtual memory (bytes)=2742329344Shuffle ErrorsBAD_ID=0CONNECTION=0IO_ERROR=0WRONG_LENGTH=0WRONG_MAP=0WRONG_REDUCE=0File Input Format Counters Bytes Read=25273File Output Format Counters Bytes Written=12134

查看结果

[dnn@master ~]$ hdfs dfs -ls /OutputDataTest
Found 2 items
-rw-r--r--   3 dnn supergroup          0 2021-03-12 07:12 /OutputDataTest/_SUCCESS
-rw-r--r--   3 dnn supergroup      12134 2021-03-12 07:12 /OutputDataTest/part-r-00000

_SUCCESS 表示运行成功
结果文件是 part-r-00000

hdfs dfs -cat /OutputDataTest/part-r-00000 查看结果

[dnn@master ~]$ hdfs dfs -cat /OutputDataTest/part-r-00000
"AS    3
"License");   3
"log   1
#   466
##  32
### 53
#export 14
$HADOOP_YARN_HOME/share/hadoop/yarn/yarn-service-examples   1
$USER   1
${HADOOP_HOME}/logs 1
${HOME}/.hadooprc   1
'-'   1
'.'   1
'hadoop    1
'mapred    1
'yarn  1
(   1
(ASF)   3
(BUT    1
(Java   2
(Note   1
(command)_(subcommand)_USER.    1
(e.g.,  1
(file/dir   1
(i.e.,  2
(period)    1
(primarily) 1
(such   1
(superficially) 1
(the    3
)   1
**MUST  1
**MUST**    1
*NOT*   1
+'%Y%m%d%H%M')" 4
--config)   1
--daemon    1
-Dcom.sun.management.jmxremote.authenticate=false  2
-Dcom.sun.management.jmxremote.port=1026" 2
-Dcom.sun.management.jmxremote.ssl=false   2
-Dhadoop.security.logger=foo). 1
-Dsun.security.krb5.debug=true 1
-Dsun.security.spnego.debug"   1
-Dyarn.server.resourcemanager.appsummary.logger=INFO,RMSUMMARY"   1
-XX:+PrintGCDateStamps 2
-XX:+PrintGCDateStamps"   1
-XX:+PrintGCDetails    3
-XX:+PrintGCTimeStamps 3
-Xloggc:${HADOOP_LOG_DIR}/gc-rm.log-$(date  4
-Xms).  1
-Xmx).  1
-blah). 1
-f  1
-ls 1
-o  2
-s)}    1
.   1
..  1
... 2
.hadooprc   1
/etc/profile.d  2
/tmp    1
10  1
2.0 3
2NN 1
<-----   1
=  1
>    12

12. 关闭Hadoop

mr-jobhistory-daemon.sh stop historyserver
# 或者 mapred --daemon stop historyserver
stop-yarn.sh
stop-dfs.sh

好几天了，跟着书，终于安装成功了！

我的CSDN博客地址 https://michael.blog.csdn.net/

长按或扫码关注我的公众号（Michael阿明），一起加油、一起学习进步！