Scontrol modified to print separate error messages for job arrays with
different exit codes on the different tasks of the job array. Applies to
job suspend and resume operations.

This will set/export only specific environment variables

job elapsed time.

the i/o thread.

exit code.

Test 3.11 was failing in some configurations without this as
the CPU count in the RPC was lower than the number of nodes in
the required node list

similar wording.

information wasn't stored in accounting.

job/step completion.

bit_unfmt.

Signed-off-by: Danny Auble <da@schedmd.com>

In the job_submit plugin: Remove all slurmctld locks prior to job_submit()
being called for improved performance. If any slurmctld data structures are
read or modified, add locks directly in the plugin.

HealthCheckProgram for nodes in any other states than IDLE.#978

Unload job tables rather than windows at job end. The table unload
also unloads job tables.

Added new internal Slurm functions xmalloc_nz() and xrealloc_nz(), which do
not initialize the allocated memory to zero for improved performance.

switch/nrt - Do not explicitly unload windows for a job on termination,
only unload its table (which automatically unloads its windows).

lost should the slurmctld restart.

Correct NumCPUs count for jobs with --exclusive option.
bug 909

lost should the slurmctld restart.

Correct NumCPUs count for jobs with --exclusive option.
bug 909

This probably only happens on native Cray systems due to the
deallocation delays related to node health check. In any case,
the symptom is error message of this sort "job # dealloc of node
... bad node_offset 0 count is 0". It then fails to deallocate
the nodes GRES back for use by other jobs.
bug 973

slurm_conf_reinit.

counting as multiple nodes.

slurmstepd attempts to create it, for example left over from a previous
requeue or crash, delete it and recreate it. #961.

switch/nrt - Unload job tables (in addition to windows) in user space mode
to avoid leaking NRT.
bug 964

Fix race condition which could result in requeue if batch job exit and node
registration occur at the same time.

(From that commit)

    There was a problem when building from source where for example
    @bindir@ would resolve to ${prefix}/bin.

    This patch fixes it, based on http://www.gnu.org/software/autoconf/manual/
    autoconf-2.69/html_node/Installation-Directory-Variables.html

    It also changes opt_modulefiles_slurm to opt_modulefiles_slurm.in but
    I couldn't figure out how to get git diff to show that.

and will avoid the need for all the daemons to link to libhwloc

Added SchedulerParameters options of bf_max_job_array_resv to control how
many tasks of a job array should have resources reserved for them.

and should there be problems accessing the state files.