The burst_buffer/cray plugin sets some environment variables for
  use by the user. This adds those environment variables to the
  salloc command

Conflicts:
	src/slurmctld/reservation.c

If an advanced reservation contains no nodes (e.g. licenses or
  burst buffers only), then avoid logic used to replace failed
  or failing nodes which would reference a node bitmap that is
  NULL, which would result in an abort

Correct some logic for limit enforcement, including change for
  emulated Cray system and change some variable names for better
  clarity

Added flag of EmulateCray and dw_wlm_cli script

fix non-system munge builds - ref bug 1842

and a node is older it will stop communicating completely.

Hooray for test21.21!

TRESBillableUnits

respect to TRESBillingUnits

a nice little segfault.

This picked up recent changes in configure.ac

Run pre_run after persistent buffer creation per Cray documenation

Add logic to burst_buffer/cray and generic to set job's buffer size
  in TRES array for limits and accounting. This includes only the
  the job-specific buffers, not persistent buffers, which will be
  handled separately.

The problem is that with high rate profiling, the cpu usage becomes a
binary field (see the output below). I would like to have a more accurate
value of the cpu usage. I'm thinking about writing another profiling
plugin that may be very useful in our environment, and probably also in
other environments where compute nodes are shared.

Bug 1748

change is to ease future transition to IPv6 reducing the number of
places that use AF_INET.  AF_SLURM currently points to AF_INET though.

char in the tres string.

removed, so make sure we process correctly, sort it while we are at it.