Sun Grid Engine Monitor


Important Note:

Current release of GEMonitor supports Sun Grid Engine 5.3 ONLY. Version 6.0 has different output (so I am told) so it will need to be reworked. That is in the process. No release date yet.

Introduction

We are currently in the process of migrating from LSF to SGE so we are trying to make people's life easier by providing them with a nicer way to monitor jobs other than the standard qstat/qhost commands. I downloaded Grid Engine portal but the installation involved was too much and besides it was an overkill.

Since I like PHP I decided to write a web front-end to qstat. The total code currently is about 500 lines of code out of which a third is a date conversion library I downloaded off the net. Below is a sample page of output. Please notice that time in queue or runtime is actually calculated so you will know if a job has been running for too long. Also some of the rows are color-coded depending on certain metrics ie. light blue if job is a sleep job (used for grabbing a token for use of particular lab setup), gray for suspended or pending jobs, red for jobs where machine load is close to 0 indicating a problem with the job.

ChangeLog

Version 0.9

Version 0.8

Download

If you would like to download the source you can get it by clicking on the following link gemonitor-0.9.tar.gz. Old releases are here.

Documentation

Installation document is here.

 SgeMonitor screenshot

Please contact me at vuksan-php@veus.hr