Am 30.01.2013 um 20:17 schrieb [email protected]: > > We're running SGE 6.2u5 and we've got a "short" queue, which assigns a > higher priority but imposes run-time and CPU-time limits. > > We also have a "short-on-interactive" queue to allow short jobs to > run on a subset of the slots on our interactive nodes, with the goal > of allowing short, high-priority jobs to use idle resources on the > interactive machines. The "short-on-interactive" queue is subordinate > to the "interactive" queue, so if the interactive server becomes busy, > the batch jobs will be suspended. > > In general, this works fine. However, if a batch job is suspended for too > long, it exceeds the h_rt limit and is killed by SGE.
Yep. It's more wall-clock than run-time. > Is there any way to prevent a job from accumulating "run time" in SGE's > accounting for the period that it is suspended? Not that I'm aware of. > If this is not possible now, > can this be considered as a future request for enhancement? This would mean to distinguish between wall-clock and granted run-time for execution. Using h_cpu is no option for now as your are oversubscribing the machines? -- Reuti _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
