Page History

...

yee's two points:
(1) need to limit the number of non-preemptable jobs, in particular the
    single-core jobs which can scatter across nodes
(2) how do we preempt the jobs and limit the resources used by preemptable
    jobs

yee's proposal:
(1) limit every facilities use of non-preemptable jobs to what
    they have purchased (e.g. LCLS is limited to 88 nodes of
    non-preemptable jobs, or half of 176 nodes)
    - coact can help with this
    - is this a hard limit, or does it have a long time-constant?
      o a hard limit for a repo (enforced by slurm), but not at the
        multiple-repo level
        (multi-repo is not enforced by slurm, but coact so a long
     time-constant to see if we've crossed the 88 node
     threshold)
(2) LCLS defines the order of preemption of experiment repos
(add-on) could extend this to support cross-facility preemption

...