scheduling decisions upon the node's configuration as specified in
    slurm.conf and ignore the node's actual hardware configuration. This
    can be useful for testing.

    a job's failure.

    previously reported only reason waiting to run). Requires cold-start of
    slurmctld (-c option).

tested logic on bluegene system, seems to work fine, haven't tested on a system that allows passthroughs, should work, but haven't tested yet.

Includes Mark Grondona's addition of the slurm_spank_local_user_init()
function, with my massaging to make it more compatible with the trunk code.
However, there is no support in the trunk API task launch code (and therefore
slaunch) for slurm_spank_local_user_init().  I will add that in a future
commit.

removed proctrack from jobacct since it won't allow us to keep track of tasks. (right now we may loose a child that reparents to init)

    it's job steps that lack switch windows. Needed for LSF operation.
    Based upon slurm.hp.node_fail.patch.

    if their count exceeds the task count. Based on slurm.hp.srun.patch.

    http://www.open-mpi.org/software/plpa).