On Wed, 6 Aug 2008, Steffen Grunewald wrote:
And BTW, what also would be very useful: a user-imposed limit on simultaneously running jobs from the same cluster ("I want to be nice to my colleagues, n slots are enough even if there are more available - because I know other's work ismore important and I hate preemption")
You can essentially do this with DAGMan. Just make a DAG file with no dependencies, and use the -maxjobs or -maxidle throttles. Note that for -maxjobs, each cluster counts as one job, while for -maxidle each individual job within a cluster counts as one job. (Yes, we know these should be changed to be consistent.)
Kent Wenger Condor Team