<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>Those would be considered separate for each job.</p>
    <p>You may want to have your prolog check to see if there is an
      epilogue running and wait for the epilogue to be done before
      starting its prolog work.</p>
    <p>Brian Andrus<br>
    </p>
    On 9/27/2021 9:15 AM, Joe Teumer wrote:<br>
    <blockquote type="cite"
cite="mid:CAJDZVfvMb6dkwL_1Wxbzu81Cj0ayJ7_6mRT_mObfXy=JAg4xng@mail.gmail.com">
      <meta http-equiv="content-type" content="text/html; charset=UTF-8">
      <div dir="ltr">
        <div dir="ltr">
          <div dir="ltr">
            <div dir="ltr">
              <div dir="ltr">
                <div>Should the Prologslurmctld script only run after
                  the Epilogslurmctld script finishes?</div>
                <div><br>
                </div>
                <div>Below you can see JobA runs and completes.</div>
                <div>While Epilogslurmctld (from JobA Node A) is
                  executing on the Slurm controller the Prologslurmctld
                  script for the next job (from Job B Node A) is also
                  running on the Slurm controller.</div>
                <div><br>
                </div>
                <div>This breaks our workflow as we are expecting the
                  next Prologslurmctld script to only run when the prior
                  job is 100% completed (initial
                  Prologslurmctld completes AND Job completes AND Job
                  Epilogslurmctld completes).</div>
                <div><br>
                </div>
                <div>Prologslurmctld > Here Prolog is starting for
                  Job B (ID 812)</div>
                <div><b>2021-09-27 15:42:58,746</b> | INFO | Starting...</div>
                <div><br>
                </div>
                <div>
                  <div>Epilogslurmctld > Here Epilog is starting and
                    ending for Job A (ID 811)</div>
                  <div><b>2021-09-27 15:42:56,694</b> | INFO |
                    Starting...<br>
                  </div>
                </div>
                <div><b>2021-09-27 15:43:01,756</b> | INFO | Exiting 0
                  after main<br>
                </div>
                <div><br>
                </div>
                <div>
                  <div>[2021-09-27T15:42:50.224] debug:  sched/backfill:
                    _attempt_backfill: beginning</div>
                  <div>[2021-09-27T15:42:50.224] debug:  sched/backfill:
                    _attempt_backfill: 1 jobs to backfill</div>
                  <div>[2021-09-27T15:42:56.653] _job_complete:
                    JobId=811 WEXITSTATUS 0</div>
                  <div>[2021-09-27T15:42:56.653] debug:  email msg to
                    root: Slurm Job_id=811
                    Name=JobA_BIOS_fixedfreq_1067mclk_nps1.ini Ended,
                    Run time 00:22:36, COMPLETED, ExitCode 0</div>
                  <div><b>[2021-09-27T15:42:56.657]</b> _job_complete:
                    JobId=811 done</div>
                  <div>[2021-09-27T15:42:58.703] debug:  sched: Running
                    job scheduler for full queue.</div>
                  <div><b>[2021-09-27T15:42:58.704]</b> debug:  email
                    msg to root: Slurm Job_id=812
                    Name=JobA_BIOS_fixedfreq_1600mclk_nps1.ini Began,
                    Queued time 00:22:38</div>
                  <div>[2021-09-27T15:42:58.704] sched: Allocate
                    JobId=812 NodeList=xxx #CPUs=256 Partition=xxx</div>
                </div>
              </div>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
  </body>
</html>