Why is my job taking a long time to start on Cheaha?

Locales: Data Centers and Campus RC University of Alabama at Birmingham
1

My job has been in queue for a long time.

–OR–

My Open OnDemand interactive job has been waiting to start for a long time.

Why is my job taking so long to start, and what can I do about it?

2

There are a few common reasons why a job can take a long time to start. These apply to jobs on the Open OnDemand web portal, and at the terminal using sbatch.

  • There is a long queue wait time. If possible, you can move ahead in the queue by canceling and resubmitting your job with fewer requested resources, or by canceling other jobs you have running that you may not need. Please use squeue -u $USER to review your running jobs.

  • Your job is requesting more resources than are available or allowed. Double check our resource limits for various situations:

    • Requesting many jobs at once? Our Quality of Service Limits docs table lists global quotas for each partition. When these limits are reached, no more jobs will be started until some resources are freed from other jobs ending.
    • Requesting jobs on a single node? Our Cheaha HPC Cluster docs table lists resource amounts for each node type. These are physical limits of available hardware, and can’t be exceeded.

  • There is a system outage. If this is the case, information will be distributed by Research Computing.