Hello all.
We have a 33 nodes cluster composed of 9 4-cores and 24 8-cores nodes. We have partitioned so that the 4-nodes are together and 8-nodes work in a separate queue. One of our users sent yesterday a job for 128 tasks : #PBS -l nodes=11:ppn=8+10:ppn=4 MAUI refused to run it, telling : multi-req PBS jobs not allowed I've found in torqueusers mailing list, that I should put : ENABLEMULTIREQJOBS TRUE into my maui.cfg file... now it is done, and the job does not start for another (unknown) reason... checkjob says : checking job 6393 State: Idle Creds: user:foury group:mab class:bonobo3j qos:DEFAULT WallTime: 00:00:00 of 3:00:00:00 SubmitTime: Wed Jul 29 18:56:44 (Time Queued Total: 00:00:01 Eligible: 00:00:01) StartDate: 00:00:01 Wed Jul 29 18:56:46 Total Tasks: 16 Req[0] TaskCount: 16 Partition: ALL Network: [NONE] Memory >= 0 Disk >= 0 Swap >= 0 Opsys: [NONE] Arch: [NONE] Features: [bonobo] Dedicated Resources Per Task: PROCS: 1 MEM: 1920M Req[1] TaskCount: 20 Partition: ALL Network: [NONE] Memory >= 0 Disk >= 0 Swap >= 0 Opsys: [NONE] Arch: [NONE] Features: [NONE] Dedicated Resources Per Task: PROCS: 1 MEM: 1920M IWD: [NONE] Executable: [NONE] Bypass: 0 StartCount: 0 PartitionMask: [ALL] Flags: RESTARTABLE Reservation '6393' (00:00:01 -> 3:00:00:01 Duration: 3:00:00:00) PE: 36.00 StartPriority: 1 cannot select job 6393 for partition DEFAULT (startdate in '00:00:01') I wonder why I get this startdate... My maui.cfg file contains (only lines changed from the default) : BACKFILLPOLICY FIRSTFIT RESERVATIONPOLICY CURRENTHIGHEST NODEALLOCATIONPOLICY CPULOAD JOBNODEMATCHPOLICY EXACTNODE RESOURCELIMITPOLICY MEM:ALWAYS:SUSPEND NODEACCESSPOLICY SHARED ENABLEMULTIREQJOBS TRUE What do I miss ??? -- Jacques Foury administrateur systemes, reseaux, clusters Institut de Mathematiques de Bordeaux http://www.math.u-bordeaux1.fr/maths/cellule _______________________________________________ mauiusers mailing list [email protected] http://www.supercluster.org/mailman/listinfo/mauiusers
