实验环境中搭建hadoop集群（3台示例）
1.修改主机名、修改hosts
vi /etc/sysconfig/network，然后将HOSTNAME修改成hadoop-node1（自定义）
vi /etc/hosts ，添加hostname及其对应的ip
2.安装jdk和配置环境变量
2.1jdk下载安装不赘述
2.2环境变量：
/etc/profile文件中添加：
export HADOOP_HOME=/home/scada/hadoop（hadoop安装路径） export PATH=$PATH:$HADOOP_HOME/bin export PATH=$PATH:$HADOOP_HOME/sbin export HADOOP_CONF_DIR=$HADOOP_HOME/etc/hadoop export YARN_CONF_DIR=$HADOOP_HOME/etc/hadoop export JAVA_HOME=/home/scada/jdk1.7.0_80（jdk安装路径） export PATH=$PATH:$JAVA_HOME/bin3.实现主节点可无秘钥登录其他节点
3.1安装ssh（ps -e | grep ssh存在ssh进程可忽略）
sudo apt-get install openssh-server
3.2在主节点上生成秘钥对
ssh-keygen -t rsa（一直enter即可）
3.3将公钥加入authorized_keys
cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys
3.4将主节点上的authorized_keys拷贝到其他主机上
3.5修改所有节点上authorized_keys权限：
chmod 600 .ssh/authorized_keys
4.解压hadoop包，在hadoop目录下创建文件系统目录
tar -zxvf hadoop...
mv hadoop... hadoop
cd hadoop
mkdir /dfs/name -p
mkdir /dfs/data -p
mkdir /tmp
5.配置相关hadoop文件
5.1 hadoop-env.sh, yarn-env.sh
修改JAVA_HOME值：export JAVA_HOME=/home/scadajdk1.7.0_80（jdk安装路径）
5.2 slaves（集群中从节点HOSTNAME）
hadoop-node2 hadoop-node35.3 core-site.xml
<configuration> <property> <name>fs.defaultFS</name> <value>hdfs://hadoop-node1:9000</value> </property> <property> <name>hadoop.tmp.dir</name> <value>file:/home/scada/hadoop/tmp</value> </property> </configuration>5.4 hdfs-site.xml
<configuration> <property> <name>dfs.namenode.secondary.http-address</name> <value>hadoop-node1:9001</value> </property> <property> <name>dfs.namenode.name.dir</name> <value>file:/home/scada/hadoop/dfs/name</value> </property> <property> <name>dfs.datanode.data.dir</name> <value>file:/home/scada/hadoop/dfs/data</value> </property> <property> <name>dfs.replication</name> <value>2</value> </property> </configuration>5.5 mapred-site.xml
<configuration> <property> <name>mapreduce.framework.name</name> <value>yarn</value> </property> </configuration>5.6 yarn-site.xml
<configuration> <property> <name>yarn.resourcemanager.hostname</name> <value>hadoop-node1</value> </property> <property> <name>yarn.nodemanager.aux-services</name> <value>mapreduce_shuffle</value> </property> </configuration>
6.将hadoop目录拷贝到其他节点
7.启动hadoop
hdfs namenode -format
start-all.sh
// hadoop-node1:/home/scada % jps 30099 NameNode 30270 SecondaryNameNode 22019 Jps 27124 ResourceManager
// hadoop-node2:/home/scada/hadoop/etc/hadoop % jps 17009 DataNode 30414 Jps 30277 NodeManager
测试：
1.上传文件：
/home/scada % hadoop fs -put jdk-7u80-linux-x64.tar.gz / put: Call From hadoop-node1/172.16.1.216 to hadoop-node1:9000 failed on connection exception: java.net.ConnectException: 拒绝连接; For more details see: http://wiki.apache.org/hadoop/ConnectionRefuse上传文件失败，查看datanode节点日志：
tail yarn-scada-nodemanager-hadoop-node2.log 2017-04-01 06:48:00,878 INFO org.apache.hadoop.ipc.Client: Retrying connect to server: hadoop-node1/172.16.1.216:8031. Already tried 8 time(s); retry policy is RetryUpToMaximumCountWithFixedSleep(maxRetries=10, sleepTime=1000 MILLISECONDS) 2017-04-01 06:48:01,879 INFO org.apache.hadoop.ipc.Client: Retrying connect to server: hadoop-node1/172.16.1.216:8031. Already tried 9 time(s); retry policy is RetryUpToMaximumCountWithFixedSleep(maxRetries=10, sleepTime=1000 MILLISECONDS)问题解决方法：http://www.cnblogs.com/dyllove98/archive/2013/06/20/3147024.html
上传文件：
// hadoop-node1:/home/scada % hadoop fs -put jdk-7u80-linux-x64.tar.gz /
查看是否上传成功：
// hadoop-node1:/home/scada % hadoop fs -ls / Found 4 items drwxr-xr-x - scada supergroup 0 2017-03-27 08:35 /data -rw-r--r-- 2 scada supergroup 153530841 2017-04-01 06:57 /jdk-7u80-linux-x64.tar.gz drwxr-xr-x - scada supergroup 0 2017-03-28 00:59 /test drwx------ - scada supergroup 0 2017-03-27 08:35 /tmp
将本地目录下的jdk压缩包删除，从hadoop分布式文件系统中下载刚刚上传的文件
// hadoop-node1:/home/scada % ls jdk-7u80-linux-x64.tar.gz // hadoop-node1:/home/scada % rm jdk-7u80-linux-x64.tar.gz // hadoop-node1:/home/scada % ls // hadoop-node1:/home/scada % hadoop fs -get /jdk-7u80-linux-x64.tar.gz // hadoop-node1:/home/scada % ls jdk-7u80-linux-x64.tar.gzok，先到这了~