information wasn't stored in accounting.

This is a set of tools and utilities to augment Slurm

job/step completion.

bit_unfmt.

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

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.

lost should the slurmctld restart.

uses SelectType=select/cons_res otherwise don't send the
request to slurmctld and return an error.

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

Test for NULL array pointers before trying to free the array elements.
We are not seeing problems with this, but just hardening the code.

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 should there be problems accessing the state files.

Note that the map/mask specified applies to all allocated nodes.

Fix for possible abort on change in GRES configuration.
bug 958