Hadoop multi-node config for ubuntu

Steps

1. hadoop user
2. SSH
3. config JAVA_HOME, HADOOP_HOME
4. config hadoop_env.sh and *-site.xml
5. config master and slaves ip addresses using /etc/hosts
6. cleanup original data on each server and format NameNode
7. start the deamons using start-*.sh

NOTICE:

• names in slaves file should be domain names
• the domain name must not contain _!

Web Interfaces

• NameNode - http://localhost:50070/
• JobTracker - http://localhost:50030/
• TaskTracker - http://localhost:50060/

Tutorials

• single-node
• multi-node
• config file simple description

Other Solutions(made by Enterprise)

• hortonworks
• cloudera - QuickStarts, Cloudera Manager, Cloudera Director