The first step is to select the tar.gz of the Hadoop version you want to install and extract the compressed files to the specified directory.
The second step, create a folder to hold the data, the name of this folder can be self-command, but to include three sub-folders (these three subfolders, can be separated, but generally we put them in the same folder)
Of these three folders, where data (the Datanode node is used, the contents of the storage is saved) is 755, the other is 777.
The third step, set the parameters
Set hadoop-env.sh content
Set Core-site.xml content
Set Hdfs-site.xml file contents
Set Mapred-site.xml content
Set Masters File contents
Set Slaves file contents
Fourth step, format Hadoop namenode
Execute./bin/hadoop Namenode-format, enter y during execution, and see success for success.
Fifth step, start the Hadoop cluster
./bin/start-all.sh
After start-up through the JPS can be found,
Among them see Datanode, Secondarynamenode, NameNode, Tasktracker, Jobtracker five.
You can also view the startup situation through a Web page:
Finally, close the cluster
stop-all.sh
hadoop--installation 1.2. Version 1