AWS Thinkbox Discussion Forums

Ignoring Priority

It seems Deadline 10 is ignoring priority

The job with lower priority ran for 45 minutes and rendered dozens of frames and never pulled off the job to move to the newer, higher priority Job.

Can you confirm that there is no other reason why it would never start rendering?

  • Is the scheduling algorithm still the default “Pool, Priority, First-In First Out” ?
  • When you suspend the 50 Priority job, does the 51 Priority job actually start rendering?
  • When you select the 51 Priority job and check the Job Candidates Filter option in the Slaves panel, do you see any Slaves listed?

If the answer of all these questions is Yes, we will have to look deeper.

We spent like 2 years trying to track this down with Thinkbox. I think the problem was corrupt entries in our database. Purging everything and starting over from scratch for groups/slaves/limits etc seems to have fixed it.

I’ve made a video demonstrating the problem. Here is what the video shows:

  1. Deadline iss set to Pools, Priority, First in First out
  2. Submit a job with 100 priority (suspended)
  3. Submit a job with 1 priority (suspended)
  4. Resume the 1 priority job and let it get started.
  5. Verify that the Slaves are permited to render both jobs.
  6. Resume 100 priority job.

Slaves continue to render current job and do not move, even after completing frames

  1. Suspend 1 priority job and confirm that slaves will then move on to 100 priiority job and render successfully.

Slave rendering was spead up 500% to save you watching time.

dropbox.com/s/27jk820lo38ag … .mp4?raw=1

Privacy | Site terms | Cookie preferences