[GE users] Advance reservations limited to s_rt=20:20:0 ?

kisielk kamil at kamilkisiel.net
Fri Oct 22 00:27:58 BST 2010


I'm seeing some weird behavior with advance reservations and limiting the walltime for jobs. See the output of the commands below:

umeshu:~ kamil$  qrsub -l s_rt=100:0:0 -d 500:0:0
Your advance reservation 58 has been granted

umeshu:~ kamil$  qrsh -now no -ar 58 -l s_rt=100:0:0 -w v
Job 332508 cannot run in queue instance "maintenance.q" because it was not reserved by advance reservation 58
Job 332508 cannot run in queue instance "interactive.q" because it was not reserved by advance reservation 58
Job 332508 (-l s_rt=360000) cannot run in queue "batch.q at node080.example.com" because it offers only qf:s_rt=20:20:00
verification: no suitable queues

umeshu:~ kamil$ qstat -q batch.q at node080.example.com -F
queuename                      qtype resv/used/tot. load_avg arch          states
---------------------------------------------------------------------------------
batch.q at node080.example.com BP    1/7/8          5.54     lx26-amd64    
        gc:high_io=400
        hl:arch=lx26-amd64
        hl:num_proc=8
        hl:mem_total=23.530G
        hl:swap_total=0.000
        hl:virtual_total=23.530G
        hl:load_avg=5.540000
        hl:load_short=5.620000
        hl:load_medium=5.540000
        hl:load_long=5.350000
        hl:mem_free=21.631G
        hl:swap_free=0.000
        hl:virtual_free=21.631G
        hl:mem_used=1.899G
        hl:swap_used=0.000
        hl:virtual_used=1.899G
        hl:cpu=70.800000
        hl:m_topology=SCCCCSCCCC
        hl:m_topology_inuse=SCCCCSCCCC
        hl:m_socket=0
        hl:m_core=0
        hl:np_load_avg=0.692500
        hl:np_load_short=0.702500
        hl:np_load_medium=0.692500
        hl:np_load_long=0.668750
        hf:infiniband=1.000000
        hc:gpu=0
        hc:h_vmem=13.000G
        hc:exclusive=1.000000
        hf:tesla_s1070=1.000000
        qf:qname=batch.q
        qf:hostname=node080.example.com
        qc:slots=1
        qf:tmpdir=/tmp
        qf:seq_no=1
        qf:rerun=0.000000
        qf:calendar=NONE
        qf:s_rt=6:23:59:00
        qf:h_rt=7:00:00:00
        qf:s_cpu=infinity
        qf:h_cpu=infinity
        qf:s_fsize=infinity
        qf:h_fsize=infinity
        qf:s_data=infinity
        qf:h_data=infinity
        qf:s_stack=infinity
        qf:h_stack=infinity
        qf:s_core=infinity
        qf:h_core=infinity
        qf:s_rss=infinity
        qf:h_rss=infinity
        qf:s_vmem=infinity
        qf:min_cpu_interval=00:05:00

umeshu:~ kamil$  qrsh -now no -ar 58 -l s_rt=20:19:00 -w v
node080:~ kamil$

The same behavior happens when using qsub, I'm just using qrsh -now no here for illustrative purposes. Where is the limit of 20:20:0 coming from?

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=288998

To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].



More information about the gridengine-users mailing list