[slurm-users] Creating a partition with memory and CPU limits

[mercan]

Hi;

I think you should set

SelectType=select/cons_res

and plus one of these:

SelectTypeParameters=CR_Memory
SelectTypeParameters=CR_Core_Memory
SelectTypeParameters=CR_CPU_Memory
SelectTypeParameters=CR_Socket_Memory

to open Memory allocation tracking according to documentation:

https://slurm.schedmd.com/cons_res_share.html

Regards;

Ahmet M.


19.07.2019 23:58 tarihinde Calvin Dodge yazdı:
> I'm trying to create a partition with memory and CPU limits, enforced 
> by cgroups. The goal is to limit jobs in the partition to 1/4 the CPUs 
> and memory available on a single node.
>
> I've created a QoS with memory limits, and then a partition which 
> specifies that QoS.  But when I run jobs in the partition, the cgroups 
> limits are just under the actual amount of memory in the node (2G), 
> rather than the lower amount I specified.
>
> I've also created an association for the partition, QoS, and my user 
> account with that limit.  Still no effect on the cgroups memory limits.
>
> What am I missing? Is this even possible?
>
> /etc/slurm/cgroup.conf
> CgroupAutomount=yes
> ConstrainCores=yes
> ConstrainRAMSpace=yes
>
> excerpts from /etc/slurm/slurm.conf
> EnforcePartLimits=yes
> ProctrackType=proctrack/cgroup
> TaskPlugin=task/cgroup
> AccountingStorageEnforce=limits,associations
>
> QoS:
>       Name Priority  GraceTime    Preempt PreemptMode               
>  Flags UsageThres UsageFactor       GrpTRES GrpTRESMins GrpTRESRunMin 
> GrpJobs GrpSubmit     GrpWall MaxTRES MaxTRESPerNode   MaxTRESMins     
> MaxWall     MaxTRESPU MaxJobsPU MaxSubmitPU     MaxTRESPA MaxJobsPA 
> MaxSubmitPA   MinTRES
> quarterte+          0   00:00:00                cluster               
>                                    1.000000       cpu=1,mem=400M
>
> Association:
>    Cluster  Account       User  Partition     Share GrpJobs       
> GrpTRES GrpSubmit     GrpWall   GrpTRESMins MaxJobs       MaxTRES 
> MaxTRESPerNode MaxSubmit     MaxWall   MaxTRESMins      QOS   Def QOS 
> GrpTRESRunMin
>      linux    science     calvin      light         1      mem=400M 
>  mem=400M         quartertest
>
>
> This is Slurm 18.08, running on an OHPC cluster with the CentOS 7.6 OS.
>
> Sincerely,
>
> Calvin Dodge