[GE users] cpu time limits

Reuti reuti at staff.uni-marburg.de
Fri May 18 13:06:34 BST 2007


Hi,

Am 18.05.2007 um 12:58 schrieb Gerd Marquardt:

>>> is it possible to define different cpu time limits per user. Now  
>>> we run SGE 6.1, and I try to solve the problem with the new  
>>> resource quota feature, but I have no success.
>>
>> how is your setup for now?
>>
>> -- Reuti
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
>> For additional commands, e-mail: users-help at gridengine.sunsource.net
>>

can you please try to make h_cpu a FORCED complex, i.e. it must be  
requested. What (I think) indeed is missing, is an option to use the  
specified value of the resource quota for the limit, if none is  
requested in the qsub command. But this would also mean, to replace  
the defined value from the queue definition as a side effect...

-- Reuti


> My setups:
>
> qconf -sq all.q
>
> qname all.q
> hostlist @allhosts
> seq_no 0
> load_thresholds np_load_avg=1.75
> suspend_thresholds NONE
> nsuspend 1
> suspend_interval 00:05:00
> priority 0
> min_cpu_interval 00:05:00
> processors UNDEFINED
> qtype BATCH INTERACTIVE
> ckpt_list check_application-level check_transparent \
> check_userdefined
> pe_list hybrid make mpi mpi2 mpi_node mpi_round mpi_test \
> mpich2_smpd mpich2_smpd_rsh mpich2_test shm
> rerun FALSE
> slots 1,[lcn09.rrzn.uni-hannover.de=4], \
> [lcn11.rrzn.uni-hannover.de=4], \
> [lcn14.rrzn.uni-hannover.de=4], \
> [lcn10.rrzn.uni-hannover.de=4], \
> [lcn08.rrzn.uni-hannover.de=4], \
> [lcn12.rrzn.uni-hannover.de=4], \
> [lcn13.rrzn.uni-hannover.de=4], \
> [lcn07.rrzn.uni-hannover.de=4], \
> [lcn06.rrzn.uni-hannover.de=4], \
> [lcn15.rrzn.uni-hannover.de=4], \
> [lcn16.rrzn.uni-hannover.de=4], \
> [lcn05.rrzn.uni-hannover.de=4], \
> [lcn01.rrzn.uni-hannover.de=4], \
> [lcn03.rrzn.uni-hannover.de=4], \
> [lcn04.rrzn.uni-hannover.de=4], \
> [lcn02.rrzn.uni-hannover.de=4]
> tmpdir /tmp
> shell /bin/ksh
> prolog /cluster/sge/cluh/prolog.sh
> epilog /cluster/sge/cluh/epilog.sh
> shell_start_mode posix_compliant
> starter_method NONE
> suspend_method /cluster/sge/cluh/suspend_mpi.sh $job_pid $job_id
> resume_method /cluster/sge/cluh/resume_mpi.sh $job_pid $job_id
> terminate_method NONE
> notify 00:00:60
> owner_list NONE
> user_lists NONE
> xuser_lists NONE
> subordinate_list NONE
> complex_values notshared=true
> projects NONE
> xprojects NONE
> calendar NONE
> initial_state default
> s_rt 120:00:00
> h_rt 120:00:00
> s_cpu 100:00:00
> h_cpu 100:00:00
> s_fsize INFINITY
> h_fsize INFINITY
> s_data INFINITY
> h_data INFINITY
> s_stack INFINITY
> h_stack INFINITY
> s_core INFINITY
> h_core INFINITY
> s_rss INFINITY
> h_rss INFINITY
> s_vmem INFINITY
> h_vmem INFINITY
>
> qconf -srqs
>
> {
> name max_slots_per_user
> description Max. Slots per User
> enabled TRUE
> limit users {*} to slots=16
> }
> {
> name cpu_limit_per_user
> description cpu_limits per user
> enabled TRUE
> limit users zzzzmq to h_cpu=216000
> limit users {*} to h_cpu=360000
> }
>
>
>
>
> -- 
> Gerd Marquardt
> RRZN / Leibniz Universitaet Hannover
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail: users-help at gridengine.sunsource.net

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list