[torqueusers] setting up multiple execution queues

Jacqueline Scoggins jscoggins at lbl.gov
Wed Feb 22 14:05:41 MST 2006


I found my problem.  Solved!!!

on the reg_16 and reg_32 I had set "max_queuable = 0" which is not what
I meant to do.  So I removed this and now things are working as
expected.

Thanks

Jackie

On Wed, 2006-02-22 at 10:44, Jacqueline Scoggins wrote:
> Hello please help!!!!
> 
> I am really confused.  I think I did it right but it is not working:
>  
> Queue Configuration
> 
> 
> Server mini
>         server_state = Active
>         scheduling = True
>         max_running = 80
>         max_user_run = 3
>         total_jobs = 0
>         state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:0 
>                            Exiting:0
>         default_queue = batch
>         log_events = 63
>         mail_from = adm
>         query_other_jobs = True
>         resources_default.cput = 01:00:00
>         resources_default.neednodes = 1
>         resources_default.nodect = 1
>         resources_default.nodes = 1
>         resources_assigned.ncpus = 0
>         resources_assigned.nodect = 0
>         scheduler_iteration = 60
>         node_ping_rate = 300
>         node_check_rate = 600
>         tcp_timeout = 120
>         default_node = 11
>         job_stat_rate = 30
>         pbs_version = torque_1.2.0p2
> 
> Queue batch
>         queue_type = Route
>         total_jobs = 0
>         state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:0 
>                           Exiting:0
>         route_destinations = reg_32,reg_16,reg
>         enabled = True
>         started = True
> 
> Queue reg_32
>         queue_type = Execution
>         max_queuable = 0
>         total_jobs = 0
>         state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:0 
>                            Exiting:0
>         max_running = 200
>         resources_min.ncpus = 32
>         resources_default.ncpus = 0
>         resources_default.walltime = 06:00:00
>         enabled = True
>         started = True
> 
> Queue reg_16
>         queue_type = Execution
>         max_queuable = 0
>         total_jobs = 0
>         state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:0 
>                           Exiting:0
>         max_running = 200
>         resources_max.ncpus = 31
>         resources_min.ncpus = 16
>         resources_default.ncpus = 0
>         resources_default.walltime = 12:00:00
>         enabled = True
>         started = True
> 
> 
> Queue reg
>         queue_type = Execution
>         Priority = 75
>         total_jobs = 0
>         state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:0    
>                           Exiting:0
>         max_running = 20
>         resources_default.ncpus = 1
>         resources_assigned.ncpus = 0
>         resources_assigned.nodect = 0
>         max_user_run = 10
>         enabled = True
>         started = True
> 
> 
> 
> 
> *********************
> PBS Script
> *********************
> 
> #!/bin/bash
> #PBS -l ncpus=18
> echo $
> echo "Start parallel job"
> /usr/local/bin/mpirun -hostfile ~/nodes -np 18 ~/.a.out
> sleep 20
> echo "Job complete"
> 
> exit 0
> 
> 
> The job consistantly goes to the reg queue, by passing the reg_16 queue
> where I want it to go.
> 
> Here is what I want to happen:
> 
> Jobs requesting more than 16 cpus but less than 32 cpus go to reg_16
> wallclock limit of 12 hours
> Jobs requesting more than 32 cpus go to reg_32 wallclock limit of 6
> hours
> Jobs not requesting any thing goes to reg (default queue)
> 
> 
> What have I done wrong????
> 
> Thanks in advanced.
> 
> Jackie
> 
> 
> _______________________________________________
> torqueusers mailing list
> torqueusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/torqueusers



More information about the torqueusers mailing list