###Command line tools for Hadoop

Compatability - Hadoop v1.0.3

Hadoop is evolving into an eco-system comprising of different components which earlier were present in non-distributed world(Operating Systems)

In this project, we attempt to give it a UNIX/Linux like command line utility which is a very tiny step towards establishing a much friendlier interaction to the distributed environment which gives the ability to combine the goodness of nix shell with hadoop client API.

####Target user

• Cluster Admin
• Users Monitoring Jobs
• Any one who would rather stick to *nix way of doing things than the web UI interaction

###Installation and Usage

Installation and Usage Wiki Page

####Current Feature List

• Mount a cluster - Equivalent to mounting a device, get the context of a cluster
• Job Status - Equivalent to ps, list jobs and their details
• Task Status - for a given job list different kinds of tasks and their execution status(number of times failed/killed)
• Kill - kill a job or task attempt(not a task, as this is not supported by the version of hadoop being used)

####More Utils to come

• Disk Usage in Cluster - Equivalent to du

• Memory and CPU usage for a given Job

• Nodes with Frequent Job Failures

  and more..

####Got something we can include? Please raise a new issue with "Feature Request" label, add in a precise description and a use case we will pick it up from there.

Many more things to follow, keep an eye on the repo.

• Srihari Srinivasan(@systems_we_make)
• Hemanth Yamijala(@yhemanth)
• Pramod N(@machinelearner)