[slurm-dev] super slurm?

[Igor Yakushin]

Hi All,
Can slurm be used to manage (or at least to submit jobs to) several
clusters that are in different physical locations and that are already
managed by slurm or some other scheduler/resource manager?
I would think that it would be easier to script such a functionality
from scratch but my boss wanted me to ask this question anyway.
Thank you,
Igor

[slurm-dev] Slurm User Group Meeting - Early Registration Ending Soon

[Jacob Jenson]

The discounted Early Registration Option for for the 2016 Slurm User 
Group meeting will end on July 31. Please sign up now to take advantage 
of the Early Registration price.


 * https://slug2016.eventbrite.com
 * http://slurm.schedmd.com/slurm_ug_agenda.html

Regards,
Jacob

[slurm-dev] Re: MaxNodes

[Luque, N.B.]

I'm using slurm 2.3.4 and yes all of them running in different nodes.
And no, I don't get any error if I submit a job requesting 9 nodes.

Thanks for your help!
Best !
Noelia


De: Carlos Fenoy [mini...@gmail.com]
Enviado: martes, 26 de julio de 2016 12:31
Para: slurm-dev
Asunto: [Spam] [slurm-dev] Re: MaxNodes

Are all the jobs running in different nodes? What version of slurm are you 
using? If you try to submit a job requesting 9 nodes do you get an error?

On Tue, Jul 26, 2016 at 12:24 PM, Luque, N.B. 
> wrote:
Thanks a lot Kent Engström for your help.
I guess that what I wanted was GrpNodes, I set them now
$ sudo sacctmgr modify user nluque account=tcvu set GrpNodes=8
then I restart slurm as before and nothing change I still can run 12 jobs in 
different nodes.
Then I change AccountingStorageEnforce=limits,qos
And restart slurm.
But still, nothing changed
What am I missing?

Thanks a lot for your help!!!
Best regards,
 noelia

On 26 Jul 2016, at 11:26, Kent Engström 
> wrote:


"Luque, N.B." > writes:
Dear all,

I tried to set the maxnodes per user =8 in slurm but it is not working.
...
$ sudo sacctmgr modify user nluque account=tcvu set MaxNodes=8
...
It looks fine to me, but then...
I then submit 9-10 works and all of them were running, so, I don’t know what 
else to do.

Are you trying to set the maximum number of nodes *per job*? In that case,
MaxNodes seem to be the right limit, as per the sacctmgr man page:

MaxNodes=
   Maximum number of nodes each job is able to use in
   this association. [...]

Are you trying to set a maximum number of nodes for all jobs? Then
GrpNodes could be what you are looking for instead?

GrpNodes=
   Maximum number of nodes running jobs are able to be allocated in
   aggregate for this association and all associations which are
   children of this association. [...]

Best Regards,
/ Kent Engström, NSC




--
--
Carles Fenoy=

[slurm-dev] MaxNodes

[Luque, N.B.]

Dear all,

I tried to set the maxnodes per user =8 in slurm but it is not working.
I don’t understand all the info, I think I missing up something with the 
dababase.
So, I tell you what I did so maybe someone can help me out, telling me  what 
I’m doing wrong or where should I write in order to get an answer.

I set in slurm.conf
AccountingStorageEnforce=limits
Then I did
$ sudo sacctmgr modify user nluque account=tcvu set MaxNodes=8

And then I restarted the daemon like root :
# /etc/init.d/slurm restart

When I check it like...
$ sacctmgr show association
   ClusterAccount   User  Partition Share GrpJobs GrpNodes  GrpCPUs 
GrpSubmit GrpWall  GrpCPUMins MaxJobs MaxNodes  MaxCPUs MaxSubmit 
MaxWall  MaxCPUMins  QOS   Def QOS
-- -- -- -- - ---   
- --- --- ---   - 
--- ---  -


slurm_clu+   tcvu nluque1   
 8  
normal

It looks fine to me, but then...
I then submit 9-10 works and all of them were running, so, I don’t know what 
else to do.

Could you please help me with this?

Thanks a lot for your help and time,

With kinds regards!
Noelia

[slurm-dev] Re: CPUSpecList and reservation problem

[Diego Zuccato]

Il 25/07/2016 09:24, Danny Marc Rotscher ha scritto:

> Could you please tell me, what I’m doing wrong?
I'm really not an expert, but IIUC you're just wasting a lot of CPUs for
a process that shouldn't use so many. Pinning it to one CPU could
improve performance if the user jobs you're running are actually
saturating the machine and are massively interconnected (every task
needs the results of all the others to proceed), so that having a task
that temporarily suspends the user job to do other things could have a
cascade effect on the other tasks.
But under normal workload you shouldn't see any difference.

PS: which kind of machine is that, with 64 sockets?

-- 
Diego Zuccato
Servizi Informatici
Dip. di Fisica e Astronomia (DIFA) - Università di Bologna
V.le Berti-Pichat 6/2 - 40127 Bologna - Italy
tel.: +39 051 20 95786
mail: diego.zucc...@unibo.it