but don't wait for any reply. We remove the DOWN node from the job's
bitmap. As soon as the other nodes complete the KILL_JOB RPC, the job
transistions to some COMPLETED state.

If a node is down and not responding, don't bother to send a KILL_JOB
RPC to it. If that is the only node associated with a job, don't have
that job go through a COMPLETING state. It goes directly to a COMPLETED
state. Also preserve the NO_RESPOND flag associated with a node if its
state is changed via user request (e.g. scontrol).

or SlurmUser (which has not been identified at option parsing time anyway).

out message was sent (e.g., slurmd down, msg sent to slurmd, slurmd up
and registers, msg previously sent to slurmd times out).

   base 10 (slurm/197)

 o fail launch thread when job state is no longer "LAUNCHING"

rather than letting agent go off the end of an array.

   resulted in logfile messages going to a random fd, usually stderr of
   the job.

as needed).

   condition.

   where signal is recieved after having allocated nodes but before
   real signal handlers are installed.

   to close, do not wait for job indefinitely.

 string (this is not needed).

   failed (presumably due to unkillable processes)
 o retry failed JOB_KILL rpcs

   rpc.