[torqueusers] Resources not used
Enrico Morelli
enrico.morelli at gmail.com
Fri Dec 5 05:08:53 MST 2008
Dear all,
I'm using torque 2.1.9 and maui 3.2.6p19 on a cluster with 32 processors.
The problem that I don't understand is that I've set the maximum number of
jobs = 32 for members of the projects group (see maui.cfg after), but I can
launch only 27 jobs, the other jobs are queued.
This is the showq:
ACTIVE JOBS--------------------
JOBNAME USERNAME STATE PROC REMAINING
STARTTIME
.
.
11998 hemeup Running 1 99:22:00:43 Fri Dec 5
11:03:03
12001 hemeup Running 1 99:22:19:17 Fri Dec 5
11:21:37
12002 hemeup Running 1 99:22:23:49 Fri Dec 5
11:26:09
27 Active Jobs 27 of 32 Processors Active (84.38%)
4 of 4 Nodes Active (100.00%)
IDLE JOBS----------------------
JOBNAME USERNAME STATE PROC WCLIMIT
QUEUETIME
.
.
.
12036 prodoc Idle 1 4:00:00:00 Fri Dec 5
12:56:57
12037 prodoc Idle 1 4:00:00:00 Fri Dec 5
12:56:57
9 Idle Jobs
BLOCKED JOBS----------------
JOBNAME USERNAME STATE PROC WCLIMIT
QUEUETIME
Total Jobs: 36 Active Jobs: 27 Idle Jobs: 9 Blocked Jobs: 0
This is the maui.cfg (the jobs are submitted using a projects group member):
RMPOLLINTERVAL 00:00:10
DEFERTIME 00:01:00
#############
CLASSWEIGHT 1
CREDWEIGHT 1
USERWEIGHT 1
GROUPWEIGHT 1
SERVWEIGHT 1
QUEUETIMEWEIGHT 10
XFACTORWEIGHT 3
XFWEIGHT 7
XFCAP 1000000
ENABLEMULTIREQJOBS TRUE
JOBPRIOACCRUALPOLICY QUEUEPOLICY
JOBMAXSTARTTIME 01:00:00
#############
MAXJOBPERGROUPPOLICY ON
SMAXJOBPERGROUPCOUNT 32
MAXJOBPERGROUPCOUNT 32
MAXJOBQUEUEDPERUSERPOLICY ON
MAXJOBQUEUEDPERUSERCOUNT 5
MAXJOBQUEUEDPERGROUPPOLICY ON
MAXJOBQUEUEDPERGROUPCOUNT 10
SHORTPOOLPOLICY ON
SHORTPOOLMAXTIME 3600
SHORTPOOLMINSIZE 1
SHORTPOOLMINPCT 5
GROUPCFG[projects] PRIORITY=10000 MAXPROC=32 MAXJOB=32,32 MAXJOBQUEUED=32
CLASSCFG[cert] PRIORITY=10000 MAXJOB=2,40
GROUPCFG[enmr] PRIORITY=1000 MAXPROC=32 MAXJOB=28,28 MAXJOBQUEUED=16
CLASSCFG[short] PRIORITY=600 MAXJOB=24,28
CLASSCFG[medium] PRIORITY=400 MAXJOB=14,16
CLASSCFG[long] PRIORITY=200 MAXJOB=6,8
CLASSCFG[verylong] PRIORITY=50 MAXJOB=1,2
ENABLEMULTIREQJOBS TRUE
Where is the wrong parameter?
Thanks
Enrico Morelli
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20081205/6692f3de/attachment-0001.html
More information about the torqueusers
mailing list