Print an error rather than setting CUDA_VISIBLE_DEVICES environment
variable to  "NoDevFiles" if no device files defined.

Patch 4f68cde5bd6b4fcf839f6694457373c81d9548ba from chaos/slurm by Don Lipari, LLNL

The change in function call order of commit e60abe43
resulted in slurmd daemons on front-end systems not registering with the
proper node name.

Backfill scheduling - Add SchedulerParameters configuration parameter of
"bf_res" to control the resolution in the backfill scheduler's data about
when jobs begin and end. Default value is 60 seconds (used to be 1 second).

Do not drain the compute or front-end node when trying to start a job
for which the UID is not found

Release locks on cray system after inventory and before backfill scheduling loop
in order to not process more jobs and avoid blocking pending RPCs for so long

Add configure option of "--without-rpath" which builds SLURM tools without
the rpath option, which will work if Munge and BlueGene libraries are in
the default library search path and make system updates easier.

This patch contains corrections for spelling errors in the code and
improvements for some man pages. Patch from Gennaro Oliva.

Patch by Stephen Trofinoff, CSCS.

Change the logging of a job step create failure due to SlurmctldProlog
running from info() to debug() since this can be due to a race condition.

Do not attempt to run HeathCheckProgram on powered down nodes. Patch from
Ramiro Alba, Centre Tecnològic de Tranferència de Calor, Spain.

This reverts https://github.com/hautreux/slurm/commit/35075c10995f4e83d0104662f147cd7b413d25f4
My version of RPM doesn't seem to understand the "replace"
parameter for %config, so rpm builds bomb out with
    Invalid %config token: replace
    Invalid %config token: replace

I think the default for %config is to not replace config
files, so I'm not sure this commit is needed. Since the
release agents are not really config files, but need to
be tightly coupled to the release of task/cgroup and
proctrack/cgroup, I would say the release agents should
be moved under /usr/libexec/slurm

At the very least, this commit needs to be reverted before slurm-2.3.1
in order to allow rpms to be built.

Thanks,
Mark Grondona

Report correct job "Reason" if needed nodes are DOWN, DRAINED, or
NOT_RESPONDING, "Resources" rather than "PartitionNodeLimit".

plugins in the slurmd.

    value for jobs.

Cgroup plugins update