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

Re: [HTCondor-users] accounting for slots/cores in draining?

is there a way to account(?) for resources, that are not available due to draining? I.e., to account draining losses to some kind of "virtual" group or so?

As far as I know, there's no explicit support for this in HTCondor.

However, I have no idea, if something like that would be reasonably possible? So for example, draining on a node for a new dynamic 8core job slot - and keep track of the accumulated time over the preceding eight 1core jobs varying end dates to now.

I don't understand what your example. What do the previous set of jobs have to do with the cost of draining? Isn't the cost of draining just the slot-weight of the slots idled by draining times the duration of that idling? (Assuming you neither preempt to drain nor run backfill jobs while draining.)

- ToddM