This was an unrealized regression from commit 0da01963.  The problem
is we were clearing the job_ptr->job_resrcs too early.  This patch fixes
it to wait until the job is actually being requeued so it does the right
thing.

not running.

MB, GB, TB, PB, or N (for nodes).
Also added a BB_SIZE_IN_NODES flag to slurm.conf

The logic was flawed to preserve BurstBufferType on slurmctld reconfig
In addition for many plugins, one could go from no value to some
value (i.e. a change in plugin) without reconfiguring. In this case,
the slurmctld logic would not match what is in slurm.conf or reported
by "scontrol show config".

Move spank_user_task() call in slurmstepd after the task_g_pre_launch()
so that the task affinity information is available to spank.

Added reporting of job array ID in burst buffer info
Added basic burst buffer scheduling logic for pending jobs

pack it this way so we will not change it in 15.08

Conflicts:
	src/slurmctld/node_scheduler.c

Bug 1202

then jobs pending with invalid dependency are going to be terminated.

Conflicts:
	src/slurmctld/job_mgr.c

This isn't that big of an issue for 14.03, but 14.11 added more to this
string which could overflow the buffer since sprintf is used instead of
snprintf.  Using xstrfmtcat fixes the issue and is easier to read code.

proper functionality.

See "scontrol show burst"