1.目前只是单机环境，namenode和datanode都在一台机器。hadoop版本选的是2.7.2，jdk选的是jdk-8u131-linux-64.rpm

2.安装jdk 

rpm -ivh jdk-8u111-linux-x64.rpm

3.安装密钥

ssh -keygen -t rsa

在root目录下会自动生成.ssh目录

4.把公钥写到authorized_keys里面

5.修改权限

6.关闭防火墙

7.解压hadoop安装包

tar zxf hadoop-2.7.2.tar.gz

8.修改 /etc/profile

#java
JAVA_HOME=/usr/java/default
export PATH=$PATH:$JAVA_HOME/bin
#hadoop
export HADOOP_HOME=/hadoop_soft/hadoop-2.7.2
export HADOOP_OPTS="$HADOOP_OPTS -Djava.library.path=/hadoop_soft/hadoop-2.7.2/lib/native"
export HADOOP_COMMON_LIB_NATIVE_DIR=$HADOOP_HOME/lib/native
export HADOOP_OPTS="-Djava.library.path=$HADOOP_HOME/lib:$HADOOP_COMMON_LIB_NATIVE_DIR"
##export LD_LIBRARY_PATH=/hadoop_soft/hadoop-2.7.2/lib/native/:$LD_LIBRARY_PATH
export PATH=$PATH:$HADOOP_HOME/bin:$HADOOP_HOME/sbin

9.修改配置文件 hadoop-2.7.2/etc/hadoop/

(1) core-site.xml  fs.defaultFS  就是namenode的节点名称和地址

 
 fs.defaultFS

hdfs://192.168.1.120:9000

 hadoop.tmp.dir

/hadoop_soft/hadoop-2.7.2/current/tmp

  fs.trash.interval

  4320 

(2)hdfs-site.xml

   dfs.namenode.name.dir 

   /hadoop_soft/hadoop-2.7.2/current/dfs/name 

   dfs.namenode.data.dir 

   /hadoop_soft/hadoop-2.7.2/current/data 
 

   dfs.replication   

   1 
 

   dfs.webhdfs.enabled 

   true 

   dfs.permissions.superusergroup 

   staff
 

   dfs.permissions.enabled 

   false 

(3). yarn-site.xml

  yarn.resourcemanager.hostname 

   192.168.1.115 

  yarn.nodemanager.aux-services 

   mapreduce_shuffle

  yarn.nodemanager.aux-services.mapreduce.shuffle.class

  org.apache.hadoop.mapred.ShuffleHandler

  yarn.resourcemanager.address

 192.168.1.120:18040

  yarn.resourcemanager.scheduler.address

 192.168.1.120:18030

  yarn.resourcemanager.resource-tracker.address

  192.168.1.120:18025

  yarn.resourcemanager.admin.address

 192.168.1.120:18141

  yarn.resourcemanager.webapp.address

 192.168.1.120:18088

yarn.log-aggregation-enable

  true

yarn.log-aggregation.retain-seconds

 86400

yarn.log-aggregation.retain-check-interval-seconds

  86400

yarn.nodemanager.remote-app-log-dir

  /tmp/logs

yarn.nodemanager.remote-app-log-dir-suffix

  logs

(4).复制mapred-site.xml.template到mapred-site.xml

 
 mapreduce.framework.name

  yarn

mapreduce.jobtracker.http.address

 192.168.1.120:50030

mapreduce.jobhistory.address

  192.168.1.120:10020

 mapreduce.jobhistory.webapp.address

 192.168.1.120:19888

mapreduce.jobhistory-done-dir

 /jobhistory/done

mapreduce.intermediate-done-dir

 /jobhistory/done_intermediate

 mapreduce.job.ubertask.enable

 true

 (5).编辑slaves，添加主机的IP

192.168.1.120

(6).在hadoop-env.sh文件中添加java_home，找到文件JAVA_HOME这一行

10.格式化文件系统

Hdfs namenode –format

11.启动 hadoop-2.7.2/sbin/start-all.sh

12.验证 jps

6433 NameNode

6532 DataNode

7014 NodeManager

6762 SecondaryNameNode

6910 ResourceManager

7871 Jps

13.hadoop 基本命令

hadoop fs –mkdir /hadoop-test

hadoop fs -find / -name hadoop-test

hadoop fs -put NOTICE.txt /hadoop-test/

hadoop fs –rm -R