1, must bind the host name and IP, otherwise the remote will not connect
vi /etc/hosts

10.0.6.29  commonhost

2, set ssh free login
(1) ssh-keygen, always enter
(2)ssh-copy-id localhost

3, install Java
(1) Install jdk1.8:

yum install java-1.8.0-openjdk java-1.8.0-openjdk-devel

(2) Set environment variables:
vi /etc/profile

export JAVA_HOME=/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.181-3.b13.el7_5.x86_64/jre
export CLASSPATH=.:$JAVA_HOME/lib/rt.jar:$JAVA_HOME/lib/dt.jar:$JAVA_HOME/lib/tools.jar
export PATH=$PATH:$JAVA_HOME/bin
source /etc/profile

(3) View the installation results:
echo $JAVA_HOME
echo $CLASSPATH
echo $PATH
java -version

4、vi /etc/profile

export PATH=$PATH:$JAVA_HOME/bin:$HADOOP_HOME/bin:$HADOOP_HOME/sbin
export HADOOP_HOME=/home/hadoop

Source /etc/profile makes it effective

5. Upload and unzip hadoop-3.0.3.tar.gz

6, modify the configuration file under hadoop/etc/hadoop
(1)hadoop-env.sh：

export JAVA_HOME=/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.181-3.b13.el7_5.x86_64/jre

(2)core-site.xml：

<configuration>
    <property>
        <name>fs.defaultFS</name>
        <value>hdfs://commonhost:9000/</value>
    </property>
    <property>
        <name>hadoop.tmp.dir</name>
        <value>/home/hadoop/tmp/</value>
    </property>
</configuration>

(3)hdfs-site.xml：

<configuration>
    <property>
        <name>dfs.replication</name>
        <value>1</value>
	</property>
</configuration>

(4)mapred-site.xml：

<configuration>
    <property>
        <name>maperduce.framwork.name</name>
        <value>yarn</value>
	</property>
</configuration>

(5)yarn-site.xml：

<configuration>
     <property>
        <name>yarn.resourcemanager.hostname</name>
        <value>commonhost</value>
	</property>
    <property>
        <name>yarn.nodemanager.aux-services</name>
        <value>maperduce_shuffle</value>
    </property>
</configuration>

If you need to deploy a cluster, execute 7-10
7. Set ssh to login on the other two machines and install Java.

8, configure the master/slave under etc/hadoop
vi workers

10.0.6.29
10.0.6.39
10.0.6.49

9, remotely copy hadoop to two machines
scp -r /home/hadoop 10.0.6.39:/home/hadoop
scp -r /home/hadoop 10.0.6.49:/home/hadoop

10, remotely copy /etc/profile to two machines
scp -r /etc/profile 10.0.6.39:/etc/profile
scp -r /etc/profile 10.0.6.49:/etc/profile

11, each server bin. /hadoop namenode -format

12, start the master in the sbin directory, single machine and cluster only need to start one
./start-all.sh
jps view

13, connection test
http：//10.0.6.29:9870 //HDFS

14, MapReduce test
cd /home/hadoop/share/hadoop
hadoop jar hadoop-mapreduce-examples-3.0.3.jar pi 5 5
hadoop jar hadoop-mapreduce-examples-3.0.3.jar wordcount /work/test.txt /output