Skip to content

Latest commit

 

History

History
12 lines (7 loc) · 1.04 KB

README.md

File metadata and controls

12 lines (7 loc) · 1.04 KB

hadoop-runner

This project is used to run a hadoop MapReduce Job on a remote server, located in a private network. In this project we relay on traffic tunneling to the external server.

In order to use this project, the hadoop specific parameters should be configured first int the mk.ukim.finki.etnc.hadoop.conf.properties file. All configurations should be with the hostnames as in the example.

Next, the operating system hosts must be configured in order to resolve the hadoop masters and slaves to the loopback address. The example configuration is given in mk.ukim.finki.etnc.hadoop.hosts-patch

And the ports that need to be forwarded (for our case) are givven in the mk.ukim.finki.etnc.hadoop.tunneling-config file.

The mk.ukim.finki.etnc.hadoop.JobRunner class runs the official WordCount Map and Reduce tasks on the remote server. The Mapper and Reducer classes should be packaed in jar, and that jar should be added into the build path. This is because the JobClient should uppload the classes before it runs them.