check_jt.rb

Nagios plugin to check count of number of task trackers by parsing machines.jsp. Checked the script with different hadoop versions including CDH.

The script is written in Ruby and makes use of nokogiri gem for parsing machines.jsp.

Please install nokogiri gem before using the script.

The script has been tested using Ruby 1.9.3.

Help for running the script can be found by supplying -h argument to the script.

./check_jt.rb -h

Usage: check_jt.rb [options]
-u, –url URL URL of jobtracker
-w WARNINGNUMTTS, Warning Limit for number of TTs
–warningnumtts
-c, –criticaltts CRITICALTTS Critical Limit for number of TTs
-H, –help Display this screen

URL is the URL of the machines.jsp page. For e.g.

:

-w takes an argument which specifies minimum number of task trackers needed before warning alert is issued

-c takes an argument which specifies minimum number of task trackers needed before critical alert is issued

This script will identify blacklisted task trackers and will restart the task tracker process once again on them to bring them back online.