I determined that the problem was that the nodes were becoming over committed by the reservations that reserve two procs and 2GB of memory while also giving a specific group access to what I thought would be the rest of the resources on those nodes. Once a node is over committed by standing reservations, it will not start running any jobs. It seems maui isn't smart enough to determine the "remaining" resources on a node, i.e. if you don't specify resources in a reservation line it will try to reserve all the resources on that node regardless of resources defined in other reservation lines for the same hosts.
The solution was to use separate reservations for the various types of hosts, each with resource lines. Also the proper way to restart both pbs and maui would probably be to restart the server first, then restart maui. On Mon, Mar 22, 2010 at 7:04 PM, Sabuj Pattanayek <[email protected]> wrote: > On Fri, Mar 19, 2010 at 2:05 PM, Sabuj Pattanayek <[email protected]> wrote: >> I can't seem to get any of this to work properly. I'll post this one >> at a time to see if I can get each one to work, first I have this >> reservation: >> >> SRCFG[workinghours] >> HOSTLIST=arsenic,barium,bismuth,boron,curium,erbium,gold,helium,hydrogen,manganese,neon,niobium,nitrogen,palladium,platinum,radium,radon,selenium,silicon,silver,sulfur,tin,xenon,zinc >> SRCFG[workinghours] DAYS=[ALL] STARTTIME=12:00:00 ENDTIME=18:59:59 >> USERLIST=nobody ACCESS=DEDICATED >> _______________________________________________ mauiusers mailing list [email protected] http://www.supercluster.org/mailman/listinfo/mauiusers
