Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I would like to have two queues in a CPU/GPU cluster, one with high priority and one with low priority. Thus, jobs that are submitted in the high priority queue will be bumped to the top of the list with the waiting jobs. How is that possible in SGE?

share|improve this question

From my reading of SGE's sge_priority(5), the easiest way to configure hard priorities for jobs is to use projects to distinguish between high and low priority jobs.

Start by setting weight_priority and weight_urgency to zero, while setting weight_ticket to one. Set weight_tickets_functional to a big value (e.g. 1000000) to push functional tickets to the most important scheduling factor. This disables all user supplied and time-based priority calculations, basing all scheduling solely on the project's tickets.

Then setup projects for each priority level. Order projects by setting their fshare value. Jobs for a project with a higher fshare value are run before jobs for a project with less fshares.

You might want to use enforce_project in the main sge_conf(5) to force users to always specify a project when submitting jobs.

share|improve this answer
up vote 0 down vote accepted

So, what I did was altering the priority in a 'brute force' way, by adding a "qalter" command inside the prolog file like.

qalter -p 1000 $JOB_ID

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.