[slurm-users] ReqGRES value is not valid

[Uemoto, Tomoki]

Hi, all

I want to configure generic consumable resources(gpu) and confirm that the resources 
are assigned to jobs on each node.

I executed the following settings.

o gres.conf
  Name=gpu File=/dev/tty[0-3] CPUs=[0-24]
  Name=gpu File=/dev/tty[4-7] CPUs=[25-47]

o slurm.conf
  TaskPlugin=task/affinity,task/cgroup
　GresTypes=gpu
　NodeName=sl02-c00[1-2] Gres=gpu:2

o cgroup.conf
  CgroupAutomount=yes
  ConstrainCores=yes
  ConstrainDevices=yes

Then I submitted a sleep job with the gpu resource specified.

$ cat gresgpu.sh
#!/bin/bash

#SBATCH -J gresgpu            # Job name
#SBATCH --gres=gpu:2
#SBATCH -o job.%j.out         # Name of stdout output file (%j expands to jobId)

prun sleep 60
$

Return value of ReqGres was 0.
I don't know why it return value of ReqGres is 0.

$ sbatch gresgpu.sh
Submitted batch job 72

$ squeue -l
Fri Oct  4 14:14:36 2019
             JOBID PARTITION     NAME     USER    STATE       TIME TIME_LIMI  NODES NODELIST(REASON)
                72    normal  gresgpu     test  RUNNING       0:01 1-00:00:00      1 c001


$ sacct -j 72 -a -X -o jobid,partition,ReqGRES,AllocGRES,stat
       JobID  Partition      ReqGRES    AllocGRES      State
------------ ---------- ------------ ------------ ----------
72               normal        gpu:0        gpu:2    RUNNING
$

Regards,
Tomo