Previous logic never reopened the file, preventing proper functioning
of logrotate.

Use hash function to locate job records for improved performance.

Change partition write lock to a read lock as we use a different
mechanism for hidden partitions in getting individual jobs.

Correct logic returning remaining job dependencies in job information
reported by scontrol and squeue. Eliminates vestigial descriptors with
no job ID values (e.g. "afterany"). As depdencies are removed, the
job ID values were removed from the strings, but not the descriptors.
This eliminates both. It also checks the full job ID to make sure we do
not remove "afterany:1234" when job "123" completes.

Fix race condition on batch job termination that could result in a job exit
code of 0xfffffffe if the slurmd on node zero registers its active jobs at
the same time that slurmstepd is recording the job's exit code.
but 535

There was already cgroup locking in the version 14.03 code base
using different variable names and slighly different logic from
that in commit 3f6d9e36.
This commit is a variant of that commit in order to make the logic
in version 2.6 match that of our next release (logic which is
already pretty well tested).
bug 447

proctrack/cgroup - Add locking to prevent race condition where one job step
is ending for a user or job at the same time another job stepsis starting
and the user or job container is deleted from under the starting job step.
bug 447

This eliminates some now redundant arrays and variable copying
introduced in commit 74d1a4b4
bug 525

Substantial performance improvement for systems with Shared=YES or FORCE
and large numbers of running jobs (replace bubble sort with quick sort).
Bug 525

Remove trailing spaces
No changes in logic

Original code worked only for Cray systems. For other systems it
set gres_alloc to the total number of each GRES allocated on each
node to any job

Logs errors related to apbasil use

No change in logic, just move the logic that resets a batch job
accounting information into its own function.

No change in underlying logic

This fixes a problem where a job contains a license that is removed
in a slurmctld reconfiguration. Without this change, the job would
be left with a non-zero license_list pointer referencing memory that
had been freed
bug 527

Increase the range of possible reservation time values to allow
for a really long RPC delay (possibly due to slurmctld fail over
from primary to backup controller).
Also change to a #define value for clarity
bug 527

This might be useful for debugging, but could be easily induced
by the backfill scheduler or events introducing heavy load on
the messaging system

The time/resource allocation matrix is rebuilt on each job
exit, which severely impacts performance at large counts of
running jobs (say >10k jobs).

bug 511

This makes it simpler to enable detailed debugging for reservations.
This includes more information than we probably want to see with
the DebugFlag=reservation and would be only for developer debugging

This might have worked fine for core reservations or when there
are sufficient idle nodes to use, the the select_g_resv_test()
function clears the node bitmap for nodes that it can not use
and the reservation create logic did not restore that bitmap
after a failed resource selection attempt. This logic restores
the node bitmap on a failed call to select_g_resv_test() so we
can add nodes to the bitmap of available nodes rather than having
it repeatedly cleared.
The logic also adds some performance enhancements that I will
add to in the next commit.