tests were change appropriately. This lets one fake a Linux cluster
without all of the BGL-specific logic.

Move pretty much all BGL-specific logic into that module and the
  associated plugin and make use of an opaque data object for
  maintaining the information.

  reported as desired.
Add new function to drain node (for use by select/bluegene node
  monitoring thread).

code for current Linux clusters and develop all new logic for Blue Gene).

overwhelm the one slurmd for the entire system).

Map all nodes in cluster to a single front-end node.
Don't repeat ping/register/kill/etc. RPCs to all pseudo nodes, just the
  front-end.
Treat single message for some RPCs as representing all nodes in the
  cluster: register, ping responce, epilog complete, etc.

scontrol options. For now only the NULL plugin is available, but this
is required for ASC Purple.

option. Use SHOW_ALL as flag.

commands modified. Info filtering added to slurmctld.

flag the required state saves. Perform the saves using synchronous I/O
from just one thread. Under heavy loads this results in much faster
responsiveness and lowers slurmctld's memory and CPU overhead considerably.

    slurmctld startup
Create StateSaveLocation directory if changes via slurmctld reconfig

data structure. This eliminates risks associated with re-reading slurm.conf.

modifications were relatively minor - mostly changes in function
names or arguments.

"Scontrol abort" works. It was leaving a hung pthread due to a recent
   change.
Fix a couple of potential memory leaks
"switch_type" has been added to config data structure, un/pack, etc,
   but not yet reported to the user or documented yet.
The plugins now use function calls to get a their type and plugin directory
   from a common data structure rather than individually reading and
   parsing the configuration file.

Otherwise signal all steps associated with the job (unless individual job
steps are identified).

Allocate calls were using job table info in place after freeing mutex.
New logic copies data structures, frees mutex, sends message, and
then frees memory allocated for copy.

structure.

avoid highly fragmented resource allocations.
Add list of excluded nodes to job info dumpped and reported.
Fix how mis-matched RPC version number are handled. Let error code get
back to the API function.
Dump job state information upon each job's termination via plugin.
Re-issue incomplete write requests in job/partition state save.
Make slurmctld continue proper operation without any default partition
(gnats:317).
Add command/RPC to delete a partition.
Retry socket connection for slurmd/io.c as needed (gnats:253).

  - changed defs of HAVE_LIBELAN3 to HAVE_ELAN

node registration message to itself.

Fatal() no longer calls abort(), but terminates job using exit(1).

save on job_step completion. This should reduce the likelyhood of
information being lost on a slurmctld restart.

scalability. An arbitrary number of requests may be queued and they
are processed one per second until the queue is empty or pending
requests were last attempted recently (configuration parameters set
to 60 seconds as a minimum retry interval).

control (it needs to complete all pending RPCs and save state before
the primary reads state and takes over).

EPILOG_COMPLETE_MESSAGE. At this time the job is COMPLETED and all
associated nodes available.

returned to service. The priority is changed from 1 to value which would
be set for the job if submitted at that time.
(gnats:279)

to job_kill request and slurmctld leaves node and job in COMPLETING state
until the slurmd issues an EPILOG_COMPLETE RPC on each node. This permits
better support for non-killable processes and/or long-running epilog scripts.
Several minor changes in node registration handling and slurmctld agent
logic to better address a flood of incomming RPC (typically when system
restarts).
(gnats:268)

This prevents an orphan job if srun dies after sending the request
or the network fails or the authenticaion mechanism fails.