It may be easiest to generate such a file

JobRequeue is configured off (apply it only to node failures rather than
explicit requeue requests). Change test3.8 to match change in logic.

happening reliably today (it should be done because a 
notify_on_release flag is set in the directory).

Also made 2 the minimal job priority to distinguish the job from being held

added logic to allow for a SlurmdUser in the slurm.conf so you can allow for munge to have restrictions on signing

more nodes than configured

original access rights or UMASK are not set to good values.

We actually want to preserve the set_oomadj() function in slurm v1.4, but
have removed it completely from v1.3.

reporting the wrong value in verbose mode though (some other bug in the task/affinity plugin).

  the values of SLURMD_OOM_ADJ and SLURMSTEPD_OOM_ADJ environment
  variables. This can be used to prevent daemons being killed when
  a node's memory is exhausted.

Minor cosmetic stuff in select/cons_res

Plus some cosmetic work. Logic is somewhat different in v1.4 and v1.3.

 possible even when there are higher priority jobs ONLY in static
 partitioning mode.