[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HTCondor-users] sanity check on preemption

A job would not preempt by evicting a job, destroying partitioned slot, and creating a new partitioned slot of a different size to match the new job.

We tend to call the slot that was created as a result of dividing a partitionable slot a 'dynamic slot', but yes, you're generally right.

For some context, I enabled preemption on a cluster with partitionable slots with a PREEMPTION_REQUIREMENTS that looks if a job was running longer than 48 hours and priority is 20% higher. However, there are jobs that keep running after the 48 hour mark even when there are queued jobs waiting to run.

I assume you've verified that (a) the queued jobs match the slots on which the >48 hour jobs are running and (b) that the owner of the queued jobs in fact has a priority that's 20% higher? (How does your PREEMPTION_REQUIREMENTS determine that?)

- ToddM