[torqueusers] pbs -l procs defaults to 1 in 6.1.0 with pbs_sched?

[Joshua Moore]

Hello,

I downloaded the latest Torque, and I am trying to use

#PBS -l procs=X

I am getting

Not Running - PBS Error: Unknown node-attribute

I am using pbs_sched

I have removed all resources_default.ncpus, resources_default.nodect, and
resources_default.nodes from my queues.

I can run across 4 nodes, using #PBS -l nodes=X:ppn:Y just fine.

From this post several years ago

http://www.supercluster.org/pipermail/torqueusers/2013-December/016595.html

And the latest user manual:

http://docs.adaptivecomputing.com/torque/6-1-0/adminGuide/help.htm#topics/torque/2-jobs/requestingRes.htm?Highlight=ppn

I thought #PBS -l procs=X

should work with pbs_sched and pbs_sched should be able to schedule the
jobs across the nodes on the proper number of nodes?

Does anyone know if this works with the latest version of Torque?

Thanks in advance.

Josh

[Glen Beane]

I opened an issue for this in github in 2013. I don't think there has been
any activity on this:


https://github.com/adaptivecomputing/torque/issues/207


I'd suggest you commenton and follow this issue as well.

Post by Joshua Moore
Hello,
I downloaded the latest Torque, and I am trying to use
#PBS -l procs=X
I am getting
Not Running - PBS Error: Unknown node-attribute
I am using pbs_sched
I have removed all resources_default.ncpus, resources_default.nodect, and
resources_default.nodes from my queues.
I can run across 4 nodes, using #PBS -l nodes=X:ppn:Y just fine.
From this post several years ago
http://www.supercluster.org/pipermail/torqueusers/2013-
December/016595.html
http://docs.adaptivecomputing.com/torque/6-1-0/adminGuide/
help.htm#topics/torque/2-jobs/requestingRes.htm?Highlight=ppn
I thought #PBS -l procs=X
should work with pbs_sched and pbs_sched should be able to schedule the
jobs across the nodes on the proper number of nodes?
Does anyone know if this works with the latest version of Torque?
Thanks in advance.
Josh
_______________________________________________
torqueusers mailing list
http://www.supercluster.org/mailman/listinfo/torqueusers

[David Beer]

Just to make sure this gets a response as well -

Historically we have not updated pbs_sched. To the best of our knowledge,
very few use pbs_sched for anything other than POCs, and the overwhelming
majority use Maui or Moab for production purposes. Additionally, pbs_sched
is basically FIFO; essentially, pbs_sched gets the job list, starts at the
top, and runs jobs until it cannot run the top queued job. When resources
become available, it runs the top job, and then continues.

To put it concisely, due to very limited features and a tiny user base, I
don't really believe it's worth maintaining. If someone in the community
would like to fix this bug, we will adopt a viable pull request, but it is
not something my team anticipates working on.

Cheers,

David

Post by Glen Beane
I opened an issue for this in github in 2013. I don't think there has
https://github.com/adaptivecomputing/torque/issues/207
I'd suggest you commenton and follow this issue as well.

Post by Joshua Moore
Hello,
I downloaded the latest Torque, and I am trying to use
#PBS -l procs=X
I am getting
Not Running - PBS Error: Unknown node-attribute
I am using pbs_sched
I have removed all resources_default.ncpus, resources_default.nodect, and
resources_default.nodes from my queues.
I can run across 4 nodes, using #PBS -l nodes=X:ppn:Y just fine.
From this post several years ago
http://www.supercluster.org/pipermail/torqueusers/2013-Decem
ber/016595.html
http://docs.adaptivecomputing.com/torque/6-1-0/adminGuide/he
lp.htm#topics/torque/2-jobs/requestingRes.htm?Highlight=ppn
I thought #PBS -l procs=X
should work with pbs_sched and pbs_sched should be able to schedule the
jobs across the nodes on the proper number of nodes?
Does anyone know if this works with the latest version of Torque?
Thanks in advance.
Josh
_______________________________________________
torqueusers mailing list
http://www.supercluster.org/mailman/listinfo/torqueusers

_______________________________________________
torqueusers mailing list
http://www.supercluster.org/mailman/listinfo/torqueusers

[Joshua Moore]

I downgraded to 2.5.6 and "-l procs" works with pbs_sched.

Sometime after this version it broke.

Thanks.

Josh

Post by David Beer
Just to make sure this gets a response as well -
Historically we have not updated pbs_sched. To the best of our knowledge,
very few use pbs_sched for anything other than POCs, and the overwhelming
majority use Maui or Moab for production purposes. Additionally, pbs_sched
is basically FIFO; essentially, pbs_sched gets the job list, starts at the
top, and runs jobs until it cannot run the top queued job. When resources
become available, it runs the top job, and then continues.
To put it concisely, due to very limited features and a tiny user base, I
don't really believe it's worth maintaining. If someone in the community
would like to fix this bug, we will adopt a viable pull request, but it is
not something my team anticipates working on.
Cheers,
David

Post by Glen Beane
I opened an issue for this in github in 2013. I don't think there has
https://github.com/adaptivecomputing/torque/issues/207
I'd suggest you commenton and follow this issue as well.

Post by Joshua Moore
Hello,
I downloaded the latest Torque, and I am trying to use
#PBS -l procs=X
I am getting
Not Running - PBS Error: Unknown node-attribute
I am using pbs_sched
I have removed all resources_default.ncpus, resources_default.nodect,
and resources_default.nodes from my queues.
I can run across 4 nodes, using #PBS -l nodes=X:ppn:Y just fine.
From this post several years ago
http://www.supercluster.org/pipermail/torqueusers/2013-Decem
ber/016595.html
http://docs.adaptivecomputing.com/torque/6-1-0/adminGuide/he
lp.htm#topics/torque/2-jobs/requestingRes.htm?Highlight=ppn
I thought #PBS -l procs=X
should work with pbs_sched and pbs_sched should be able to schedule the
jobs across the nodes on the proper number of nodes?
Does anyone know if this works with the latest version of Torque?
Thanks in advance.
Josh
_______________________________________________
torqueusers mailing list
http://www.supercluster.org/mailman/listinfo/torqueusers

_______________________________________________
torqueusers mailing list
http://www.supercluster.org/mailman/listinfo/torqueusers

--
David Beer | Torque Architect
Adaptive Computing
_______________________________________________
torqueusers mailing list
http://www.supercluster.org/mailman/listinfo/torqueusers

[Joshua Moore]

BTW, I had to add resources_default.nodes=1 to all my queues. Not sure why
that is needed, but with it, pbs_sched allocates -l procs to the proper
number of cores on each node.

Post by Joshua Moore
I downgraded to 2.5.6 and "-l procs" works with pbs_sched.
Sometime after this version it broke.
Thanks.
Josh

Post by David Beer
Just to make sure this gets a response as well -
Historically we have not updated pbs_sched. To the best of our knowledge,
very few use pbs_sched for anything other than POCs, and the overwhelming
majority use Maui or Moab for production purposes. Additionally, pbs_sched
is basically FIFO; essentially, pbs_sched gets the job list, starts at the
top, and runs jobs until it cannot run the top queued job. When resources
become available, it runs the top job, and then continues.
To put it concisely, due to very limited features and a tiny user base, I
don't really believe it's worth maintaining. If someone in the community
would like to fix this bug, we will adopt a viable pull request, but it is
not something my team anticipates working on.
Cheers,
David

Post by Glen Beane
I opened an issue for this in github in 2013. I don't think there has
https://github.com/adaptivecomputing/torque/issues/207
I'd suggest you commenton and follow this issue as well.

Post by Joshua Moore
Hello,
I downloaded the latest Torque, and I am trying to use
#PBS -l procs=X
I am getting
Not Running - PBS Error: Unknown node-attribute
I am using pbs_sched
I have removed all resources_default.ncpus, resources_default.nodect,
and resources_default.nodes from my queues.
I can run across 4 nodes, using #PBS -l nodes=X:ppn:Y just fine.
From this post several years ago
http://www.supercluster.org/pipermail/torqueusers/2013-Decem
ber/016595.html
http://docs.adaptivecomputing.com/torque/6-1-0/adminGuide/he
lp.htm#topics/torque/2-jobs/requestingRes.htm?Highlight=ppn
I thought #PBS -l procs=X
should work with pbs_sched and pbs_sched should be able to schedule the
jobs across the nodes on the proper number of nodes?
Does anyone know if this works with the latest version of Torque?
Thanks in advance.
Josh
_______________________________________________
torqueusers mailing list
http://www.supercluster.org/mailman/listinfo/torqueusers

_______________________________________________
torqueusers mailing list
http://www.supercluster.org/mailman/listinfo/torqueusers

--
David Beer | Torque Architect
Adaptive Computing
_______________________________________________
torqueusers mailing list
http://www.supercluster.org/mailman/listinfo/torqueusers