<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    Hi Nicholas,<br>
    <br>
    first of all, "Happy new Year".<br>
    <br>
    I'm not sure, what you want to achieve. If you want to<br>
    a) schedule whole nodes all the time (e.g. also for serial jobs),
    your config seems on first sight ok.<br>
    b) schedule multiple serial jobs on the same node, you will have to
    replace<br>
    SelectType=select/linear<br>
    with <br>
    SelectType=select/cons_res or SelectType=select/cons_tres<br>
    You will also have to select one of the SelectTypeParameters,
    depending on your needs.<br>
    <br>
    <br>
    Best<br>
    Marcus<br>
    <br>
    <div class="moz-cite-prefix">On 1/1/20 8:49 PM, Nicholas Yue wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CADYXiJKr2adf4h-Bvmf6qMiRR8uSPLnbBftS2k+th4BhdvNNMQ@mail.gmail.com">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div dir="ltr">Hi,
        <div><br>
        </div>
        <div>  I am new to SLURM</div>
        <div><br>
        </div>
        <div>  I am experimenting with a small cluster, 1 head node
          (with GPU for visualization, not for CUDA jobs yet) and 2
          compute nodes</div>
        <div><br>
        </div>
        <div>  Each machine have quad core processors but no
          hyperthreading.</div>
        <div><br>
        </div>
        <div>  I have the following test job:</div>
        <div><br>
        </div>
        <div>---------------------------------------------------------------------</div>
        <div>#!/bin/bash<br>
          #SBATCH --job-name=producer               # Job name<br>
          #SBATCH --output=producer.%j.log          # Standard output
          and error log<br>
          #SBATCH --cpus-per-task=1<br>
          <br>
          /mnt/nfs/projects/workflow-languages/build/producer
          output.$SLURM_JOB_ID.txt<br>
        </div>
        <div>---------------------------------------------------------------------<br>
        </div>
        <div><br>
        </div>
        <div>My slurm.conf is this, is the configuration wrong ?</div>
        <div>
          <div>---------------------------------------------------------------------<br>
          </div>
          <div># slurm.conf file generated by configurator easy.html.<br>
            # Put this file on all nodes of your cluster.<br>
            # See the slurm.conf man page for more information.<br>
            #<br>
            ControlMachine=head0<br>
            ControlAddr=192.168.0.10<br>
            #<br>
            #MailProg=/bin/mail<br>
            MpiDefault=none<br>
            #MpiParams=ports=#-#<br>
            ProctrackType=proctrack/pgid<br>
            ReturnToService=1<br>
            SlurmctldPidFile=/var/run/slurm-llnl/slurmctld.pid<br>
            #SlurmctldPort=6817<br>
            SlurmdPidFile=/var/run/slurm-llnl/slurmd.pid<br>
            #SlurmdPort=6818<br>
            SlurmdSpoolDir=/var/lib/slurm-llnl/slurmd<br>
            SlurmUser=slurm<br>
            #SlurmdUser=root<br>
            StateSaveLocation=/var/lib/slurm-llnl/slurmctld<br>
            SwitchType=switch/none<br>
            TaskPlugin=task/none<br>
            #<br>
            #<br>
            # TIMERS<br>
            #KillWait=30<br>
            #MinJobAge=300<br>
            #SlurmctldTimeout=120<br>
            #SlurmdTimeout=300<br>
            #<br>
            #<br>
            # SCHEDULING<br>
            FastSchedule=1<br>
            SchedulerType=sched/backfill<br>
            #SchedulerPort=7321<br>
            SelectType=select/linear<br>
            #<br>
            #<br>
            # LOGGING AND ACCOUNTING<br>
            AccountingStorageType=accounting_storage/none<br>
            ClusterName=cluster<br>
            #JobAcctGatherFrequency=30<br>
            JobAcctGatherType=jobacct_gather/none<br>
            #SlurmctldDebug=3<br>
            SlurmctldLogFile=/var/log/slurm-llnl/slurmctld.log<br>
            #SlurmdDebug=3<br>
            SlurmdLogFile=/var/log/slurm-llnl/slurmd.log<br>
            #<br>
            #<br>
            # COMPUTE NODES<br>
            NodeName=head0 CoresPerSocket=4 ThreadsPerCore=1
            RealMemory=32768 State=UNKNOWN<br>
            NodeName=compute[0-1] CoresPerSocket=4 ThreadsPerCore=1
            RealMemory=16384 State=UNKNOWN<br>
            PartitionName=cfd Nodes=head0,compute[0-1] Default=YES
            MaxTime=INFINITE State=UP<br>
          </div>
        </div>
        <div>
          <div>---------------------------------------------------------------------<br>
          </div>
          <div><br>
          </div>
        </div>
        <div>Cheers<br>
          -- <br>
          <div dir="ltr" class="gmail_signature"
            data-smartmail="gmail_signature">
            <div dir="ltr">Nicholas Yue<br>
              Graphics - Arnold, Alembic, RenderMan, OpenGL, HDF5<br>
              Custom Dev - C++ porting, OSX, Linux, Windows<br>
              <a href="http://au.linkedin.com/in/nicholasyue"
                target="_blank" moz-do-not-send="true">http://au.linkedin.com/in/nicholasyue</a><br>
              <a href="https://vimeo.com/channels/naiadtools"
                target="_blank" moz-do-not-send="true">https://vimeo.com/channels/naiadtools</a><br>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
    <pre class="moz-signature" cols="72">-- 
Marcus Wagner, Dipl.-Inf.

IT Center
Abteilung: Systeme und Betrieb
RWTH Aachen University
Seffenter Weg 23
52074 Aachen
Tel: +49 241 80-24383
Fax: +49 241 80-624383
<a class="moz-txt-link-abbreviated" href="mailto:wagner@itc.rwth-aachen.de">wagner@itc.rwth-aachen.de</a>
<a class="moz-txt-link-abbreviated" href="http://www.itc.rwth-aachen.de">www.itc.rwth-aachen.de</a>
</pre>
  </body>
</html>