<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>Hi,</p>
    <p>I'm new to slurm (as admin) and I need some help. Testing my
      initial setup with:</p>
    <blockquote>[begou@tenibre ~]$ <b>salloc -n 1 sh</b><br>
      salloc: Granted job allocation 11<br>
      sh-4.4$ <b>squeue</b><br>
                   JOBID PARTITION     NAME     USER ST       TIME 
      NODES NODELIST(REASON)<br>
                      <b>11 </b>      all       sh    begou  R      
      0:16      1 tenibre-0-0<br>
      sh-4.4$<b> srun /usr/bin/hostname</b><br>
      srun: error: timeout waiting for task launch, started 0 of 1 tasks<br>
      srun: Job step 11.0 aborted before step completely launched.<br>
      srun: Job step aborted: Waiting up to 32 seconds for job step to
      finish.<br>
      srun: error: Timed out waiting for job step to complete<br>
    </blockquote>
    <p>I check the connections:<br>
    </p>
    <p><b>tenibre is the login node</b> (no daemon running)</p>
    <blockquote>nc -v tenibre-0-0 6818 <br>
      nc -v management1 6817<br>
    </blockquote>
    <b>management1 is the management node</b> (slurmctld running)<br>
    <blockquote>nc -v tenibre-0-0 6818<br>
    </blockquote>
    <b>tenibre-0-0 is the first compute node</b> (slurmd running)
    <blockquote>
      <p>nc -v management1 6817<br>
      </p>
    </blockquote>
    <p>All tests return "<i>Ncat: Connected...</i>"</p>
    <p>The command "id begou" works on all nodes and I can reach my home
      directory on the login node and on the compute node.</p>
    <p>On the compute node slurmd.log shows:</p>
    <blockquote>[2020-11-10T11:21:38.050]<b> launch task</b> <b>11.0 </b>request
      from UID:23455 GID:1036 HOST:172.30.1.254 PORT:42220<br>
      [2020-11-10T11:21:38.050] debug:  Checking credential with 508
      bytes of sig data<br>
      [2020-11-10T11:21:38.050] _run_prolog: run job script took usec=12<br>
      [2020-11-10T11:21:38.050] _run_prolog: prolog with lock for job 11
      ran for 0 seconds<br>
      [2020-11-10T11:21:38.053] debug:  AcctGatherEnergy NONE plugin
      loaded<br>
      [2020-11-10T11:21:38.053] debug:  AcctGatherProfile NONE plugin
      loaded<br>
      [2020-11-10T11:21:38.053] debug:  AcctGatherInterconnect NONE
      plugin loaded<br>
      [2020-11-10T11:21:38.053] debug:  AcctGatherFilesystem NONE plugin
      loaded<br>
      [2020-11-10T11:21:38.053] debug:  switch NONE plugin loaded<br>
      [2020-11-10T11:21:38.054] [11.0] debug:  Job accounting gather
      NOT_INVOKED plugin loaded<br>
      [2020-11-10T11:21:38.054] [11.0] debug:  Message thread started
      pid = 12099<br>
      [2020-11-10T11:21:38.054] debug:  task_p_slurmd_reserve_resources:
      11 0<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  task NONE plugin loaded<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  Checkpoint plugin loaded:
      checkpoint/none<br>
      [2020-11-10T11:21:38.068] [11.0] Munge credential signature plugin
      loaded<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  job_container none plugin
      loaded<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  mpi type = pmi2<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  xcgroup_instantiate:
      cgroup '/sys/fs/cgroup/freezer/slurm' already exists<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  spank: opening plugin
      stack /etc/slurm/plugstack.conf<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  mpi type = (null)<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  using mpi/pmi2<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  _setup_stepd_job_info:
      SLURM_STEP_RESV_PORTS not found in env<br>
      [2020-11-10T11:21:38.068] [11.0] debug:  mpi/pmi2: setup sockets<br>
      [2020-11-10T11:21:38.069] [11.0] debug:  mpi/pmi2: started agent
      thread<br>
      [2020-11-10T11:21:38.069] [11.0]<b> error: connect io: No route to
        host</b><br>
      [2020-11-10T11:21:38.069] [11.0] error: IO setup failed: No route
      to host<br>
      [2020-11-10T11:21:38.069] [11.0] debug: 
      step_terminate_monitor_stop signaling condition<br>
      [2020-11-10T11:21:38.069] [11.0] error: job_manager exiting
      abnormally, rc = 4021<br>
      [2020-11-10T11:21:38.069] [11.0] debug:  Sending launch resp
      rc=4021<br>
      [2020-11-10T11:21:38.069] [11.0] debug:  _send_srun_resp_msg: 0/5
      <b>failed to send msg type 6002: No route to host</b><br>
      [2020-11-10T11:21:38.169] [11.0] debug:  _send_srun_resp_msg: 1/5
      failed to send msg type 6002: No route to host<br>
      [2020-11-10T11:21:38.370] [11.0] debug:  _send_srun_resp_msg: 2/5
      failed to send msg type 6002: No route to host<br>
      [2020-11-10T11:21:38.770] [11.0] debug:  _send_srun_resp_msg: 3/5
      failed to send msg type 6002: No route to host<br>
      [2020-11-10T11:21:39.570] [11.0] debug:  _send_srun_resp_msg: 4/5
      failed to send msg type 6002: No route to host<br>
      [2020-11-10T11:21:40.370] [11.0] debug:  _send_srun_resp_msg: 5/5
      failed to send msg type 6002: No route to host<br>
      [2020-11-10T11:21:40.372] [11.0] debug:  Message thread exited<br>
      [2020-11-10T11:21:40.372] [11.0] debug:  mpi/pmi2: agent thread
      exit<br>
      [2020-11-10T11:21:40.372] [11.0] <b>done with job</b></blockquote>
    <p><br>
    </p>
    <p>But I do not understand what this "No route to host" means.</p>
    <p><br>
    </p>
    <p>Thanks for your help.<br>
    </p>
    <p>Patrick<br>
    </p>
    <p><br>
    </p>
  </body>
</html>