一. 环境准备

准备Ubuntu、JDK8、Hadoop2.8.5

其他准备

1. 更新apt sudo apt-get update

2. SSH安装，配置无密码SSH登入

2.1. SSH安装 sudo apt-get install openssh-server
2.2. 配置SSH无密码登入
cd ~/.ssh/                            # 若没有该目录，请先执行一次ssh crab2died
ssh-keygen -t rsa                     # 会有提示，都按回车就可以
cat ./id_rsa.pub >> ./authorized_keys # 加入授权
ssh crab2died                         # 验证无密码登入

3. 修改hosts

sudo vi /etc/hosts
# 添加 
本机ip   crab2died

二. 安装Hadoop

1. 解压Hadoop

cd ~
sudo tar -zxf ~/Downloads/hadoop-2.8.5.tar.gz -C /usr/local # 解压到/usr/local中
cd /usr/local/                                              
chmod -R 777 ./hadoop-2.8.5                                 # 设置权限

2. 设置Hadoop环境变量

sudo vi /etc/profile
# 添加
export HADOOP_HOME=/usr/local/hadoop-2.8.5 
export PATH=$PATH:${HADOOP_HOME}/sbin:${HADOOP_HOME}/bin
# 保存执行
source /etc/profile

3. 验证Hadoop版本

hadoop version    # 成功会返回版本信息

4. 伪分布式配置

4.1. 进入${HADOOP_HOME}/etc/hadoop目录中，修改以下文件

4.1.1. 修改 hadoop-env.sh

将export JAVA_HOME=${JAVA_HOME}改成export JAVA_HOME=/usr/local/jdk1.8.0_181 # JDK根目录

4.1.2. 修改 core-site.xml
<configuration>
    <property>
        <name>hadoop.tmp.dir</name>
        <value>file:/home/crab2died/hadoop/tmp</value>
        <description>Abase for other temporary directories.</description>
    </property>
    <property>
        <name>fs.defaultFS</name>
        <value>hdfs://crab2died:9000</value>
    </property>
</configuration>
4.1.3. 修改 hdfs-site.xml
<configuration>
   <property>
       <name>dfs.nameservices</name>
       <value>hadoop-cluster</value>
   </property>
   <property>
       <name>dfs.replication</name>
       <value>1</value>
   </property>      
   <property>
       <name>dfs.namenode.name.dir</name>
       <value>file:/home/crab2died/hadoop/hdfs/nn</value>
   </property>
   <property>
       <name>dfs.namenode.checkpoint.dir</name>
       <value>file:/home/crab2died/hadoop/hdfs/snn</value>
   </property>
   <property>
       <name>dfs.namenode.checkpoint.edits.dir</name>
       <value>file:/home/crab2died/hadoop/hdfs/snn</value>
   </property>
   <property>
       <name>dfs.datanode.data.dir</name>
       <value>file:/home/crab2died/hadoop/hdfs/dn</value>
   </property>
</configuration>
4.1.4. 先复制cp mapred-site.xml.template mapred-site.xml,再修改 mapred-site.xml
<configuration>
    <property>
         <name>mapreduce.framework.name</name>
         <value>yarn</value>
    </property>
</configuration>
4.1.5. 修改 yarn-site.xml
<configuration>
   <property>
       <name>yarn.resourcemanager.hostname</name>
       <value>crab2died</value>
   </property>
   <property>
       <name>yarn.nodemanager.aux-services</name>
       <value>mapreduce_shuffle</value>
   </property>
   <property>
       <name>yarn.nodemanager.local-dirs</name>
       <value>file:/home/crab2died/hadoop/yarn/nm</value>
   </property>
</configuration>

5. 格式化HDFS NameNode

hdfs namenode -format

6. 启动集群

6.1. 启动HDFS集群

hadoop-daemon.sh start namenode
hadoop-daemon.sh start datanode
hadoop-daemon.sh start secondarynamenode  # 伪分布式才有

6.2. 启动YARN

yarn-daemon.sh start resourcemanager
yarn-daemon.sh start nodemanager

7. jps查看进程

jps
1213 NameNOde
1261 NodeManager
1521 ResourceManager
1722 DataNode
1732 SecondrayNameNode

8. 查看HDFS管理界面

http://crab2died:50070

9. 查看YARN管理界面

http://crab2died:8088