Virtual machine-based Hadoop cluster installation
1. The software we need
Xshell, SSH secure, virtual machine, Linux centos64, Hadoop1.2.1 installation package
2. Install the above software
3, install Linux, there is no more elaboration
4. Install the JDK first
My path is
java_home=/usr/lib/jvm/jdk1.7.0_79
Path= PATH: Java_home/bin
Classpath= J AV AH OM E /LIb/d T.JaR: Java_home/lib/tools.jar
5, decompression Hadoop1.2.1
Instructions are: TAR–ZXVF hadoop-1.2.1.tar.gz
Change the name, make a soft connection
6. Configuration Files
Core-site.xml
The first one is Namenode's host number.
The second one is our TMP directory, and we are using this directory to manipulate the Datanode
And then save the exit
Modify Hdfs-site.xml
Datanode of the configuration child
Let's put our datanode in there.
Modify our Secondarynode node
The IP of our machine
7. Configure Password-free login
ssh?key g eN?Td sa?P "?F /.ssh/I dd sa Cat ~/.ssh/id_dsa.pub >> ~/.ssh/authorized_keys
Password-free login from various locations
1-"2 1-" 3 and 1-"1 2-" 2 3-"3
8. Start our program
And we have to copy the files to our datanode.
Node1 Installing the JDK
Delete the original JDK first
Configure our Environment
Modify the file inside the Etc/profile
Make environment variables effective
Node1 Installing Hadoop
Node2 Ibid.
Then there is the modification of the configuration file
Inside the hadoop-env.sh of our Hadoop configuration Java_home
configured on each node host
Start the program Dfs
"Hadoop" 8, Virtual machine-based Hadoop1.2.1 fully distributed cluster installation