[torqueusers] resources_max / resources_min
Roman Zubatyuk
zubatyuk at gmail.com
Thu Jun 12 08:20:05 MDT 2008
Hi all!
I'm new to Torque, so I'm sorry if this issue have been discussed before.
I'm using Torque/Maui on a Rocks 5.0 cluster. The problem is that
resources_max / resources_min queue attributes seems to not work
correctly.
[roman at zeon ~]$ qstat -Qf PB
Queue: PB
queue_type = Execution
max_user_queuable = 10
total_jobs = 1
state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:1 Exiting:0
max_running = 2
acl_host_enable = False
acl_hosts = compute-3-29,compute-3-19,compute-3-9,compute-3-28,compute-3-18,
compute-3-8,compute-3-27,compute-3-17,compute-3-7,compute-3-26,
compute-3-16,compute-3-6,compute-3-25,compute-3-15,compute-3-5,
compute-3-24,compute-3-14,compute-3-4,compute-3-23,compute-3-13,
compute-3-3,compute-3-22,compute-3-12,compute-3-2,compute-3-31,
compute-3-21,compute-3-11,compute-3-1,compute-3-30,compute-3-20,
compute-3-10,compute-3-0
resources_max.nodes = 16
resources_max.pmem = 1800mb
resources_max.walltime = 400:00:00
resources_min.nodes = 16
resources_default.nodes = 16
mtime = 1213180918
resources_assigned.nodect = 0
max_user_run = 1
enabled = True
started = True
[roman at zeon ~]$ qsub -q PB -I -l nodes=1
qsub: Job exceeds queue resource limits MSG=cannot satisfy queue min
nodes requirement
[roman at zeon ~]$ qsub -q PB -I -l nodes=2
qsub: waiting for job 586.zeon.xray.lab to start
qsub: job 586.zeon.xray.lab ready
---------------------------------------------------------
Job 586.zeon.xray.lab started execution on node compute-3-15.local at
Thu Jun 12 17:08:06 EEST 2008
Job was selected from queue PB
Assigned nodes:
compute-3-15
compute-3-14
Total CPUs assigned: 2
---------------------------------------------------------
[roman at compute-3-15 ~]$
So, the PB queue (I guess) should accept jobs only for 16 nodes. It
does not accept job on 1 node, but accepts on 2 nodes, and also on 17
CPUs.
Thanks for your help!
Roman Zubatyuk.
More information about the torqueusers
mailing list