hi,
there is wrapper script which submits job.
raw command: qrsh -V -pty y -cwd -noshell -verbose -N run -l mf=30000M
-l testing /bin/tcsh -fc ././run
qhost -F mem_free -h <node>
Host Resource(s): hc:mem_free=17.162G
Your job 552060 ("run") has been submitted
waiting for interactive job to be scheduled ...
Your "qrsh" request could not be scheduled, try again later.
i thought it should be in qw state till hc:mem_free is more than 30G, but
in my scenarioi it is not happening.
qmaster/messages:
02/10/2012 00:15:44|worker|master|E|unable to find job 552060 from the
scheduler order package
02/10/2012 00:15:44|worker|masterW|Skipping remaining 0 orders
02/10/2012 00:15:44|schedu|master|E|unable to find job 552060 from the
scheduler order package
02/10/2012 00:15:44|schedu|master|E|could not find job "552060" in master
list
02/10/2012 00:15:44|schedu|masterE|callback function for event "779087.
EVENT DEL JOB 552060.1" failed
On Fri, Feb 3, 2012 at 2:46 AM, Reuti <[email protected]> wrote:
> Am 02.02.2012 um 17:27 schrieb vamsi krishna:
>
> > qhost -F mem_token
> >
> > Host Resource(s): hc:mem_token=45.209G
> >
> > qhost -F mem_free
> >
> > node: lx24-amd64 8 0.00 47.2G 376.4M 64.0G 0.0
> > Host Resource(s): hc:mem_free=45.209G
> >
> > seems to be working with case2: am i correct?
>
> Yep.
>
> >
> > Regards
> > VK
> >
> > On Thu, Feb 2, 2012 at 9:40 PM, Reuti <[email protected]>
> wrote:
> > Am 02.02.2012 um 17:04 schrieb vamsi krishna:
> >
> > > Case1: qconf -sc
> > > mem_free mf MEMORY
> <= YES NO 0 0
> > > mem_token mem_token MEMORY
> <= YES YES 0 0
> > >
> > > qstat -j <jid>
> > >
> > > hard resource_list: testing=TRUE,mem_token=2G,mem_free=2G
> > >
> > > usage 1: cpu=00:00:00, mem=0.00000 GBs, io=0.00000,
> vmem=N/A, maxvmem=N/A
> > >
> > > case2: qconf -sc
> > > mem_free mf MEMORY
> <= YES YES 0 0
> > > mem_token mem_token MEMORY
> <= YES YES 0 0
> >
> > Case 2 seems suitable.
> >
> > Maybe the value of the load sensor is already below the complex values:
> >
> > $ qhost -F mem_free
> >
> > Does it list (host load):
> >
> > Host Resource(s): hl:
> >
> > or (host complex)
> >
> > Host Resource(s): hc:
> >
> > ==
> >
> > What is mem_token?!?
> >
> > -- Reuti
> >
> > PS: Please reply to all.
> >
> >
> > >
> > > load_scaling NONE
> > > complex_values
> mem_free=48294M,mem_token=48294M,OS=Linux,OSREL=RHEL5, \
> > > OSSUBREL=5u7,cpu_m=intel,testing=TRUE
> > > load_values
> arch=lx24-amd64,num_proc=8,mem_total=48294.953125M, \
> > >
> swap_total=65535.992188M,virtual_total=113830.945312M, \
> > > load_avg=0.010000,load_short=0.030000, \
> > > load_medium=0.010000,load_long=0.000000, \
> > > mem_free=47917.878906M,swap_free=65535.992188M, \
> > >
> virtual_free=113453.871094M,mem_used=377.074219M, \
> > > swap_used=0.000000M,virtual_used=377.074219M, \
> > > cpu=0.000000,m_topology=SCCCCSCCCC, \
> > > m_topology_inuse=SCCCCSCCCC,m_socket=2,m_core=8,
> \
> > > np_load_avg=0.001250,np_load_short=0.003750, \
> > > np_load_medium=0.001250,np_load_long=0.000000
> > >
> > > mem_free is not reducing.
> > >
> > > regards
> > > VK
> > >
> > >
> > > On Thu, Feb 2, 2012 at 7:54 PM, Reuti <[email protected]>
> wrote:
> > > Hi,
> > >
> > > Am 02.02.2012 um 15:08 schrieb vamsi krishna:
> > >
> > > > qsub -l mem_token=2G -l mem_free=2G vsim
> > > >
> > > > qconf -se <node> is not subtracting the mem_free with 2G, is there
> any thing should be done
> > >
> > > is it defined as "consumable yes" in `qconf -sc`?
> > >
> > > -- Reuti
> > >
> > >
> >
> >
> >
>
>
_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users
