What is the correct way to specify how long a job can run on a computer cluster using LSF as the resource manager/ job scheduler?

LSF is a job scheduler that computing clusters use to allocate resources and decide when a job that users submit gets run and how it gets run. Based on IBM's knowledge base, can I do the following if I needed 5 days?

busb -We 120:00
  • What's LSF? Please edit. – a CVn Jul 15 '16 at 14:38
  • Edited, to be honest wasn't entirely sure this i the most appropriate place to ask, but hopefully? – atomsmasher Jul 15 '16 at 14:51
  • If it's about UNIX, Unix-like OSes, or Linux, and not about programming, then it's probably fine here. See What topics can I ask about here?. I'm not familiar enough with the subject matter of this specific question to be able to give an answer to it, though. – a CVn Jul 15 '16 at 14:56

With bsub -We HH:MM you give LSF an estimate of how long your job will be running for, in hours and minutes.

In contrast to using -W, -We is not setting a hard limit but only helps LSF with sorting out the scheduling of your and other's jobs on the cluster, and your job will not be killed if it exceeds the given estimate.

There might be other limits in place that makes your job unschedulable, such as a maximum run time for the queue you're submitting your job into. It is likely that you job, in the case that you ask for too much time, gets stuck in a pend (pending) state indefinitely.

My LSF knowledge is a bit old, so I'm not entirely certain about that last paragraph.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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