Alongside previous commit, fixes errors like:

error: acct_policy_handle_accrue_time: QOS normal accrue_cnt underflow

Bug 5863.

Would have been incorrectly set to 0 if begin_time hadn't been
established yet.

Bug 5863.

a jobs --mem-per-cpu and also -c is greater than 1.

Bug 5858

PreserveCaseUser.

Bug 5813

Bug 5813

curr_inx needs to be reset correctly on each line. Overlooked
in commit cbc0b901 when that was moved to a global variable.

Bug 5831.

/proc/sys/kernel/pid_max.

Bug 5697

scheduled right away.

Bug 5774

Basically this reverts ec59bb1b which appears to be an academic
performance patch.  When testing a bitmap of 10000 with every other
bit set it only took at most 4 usec with most taking 1 usec.

of an array.

Bug 5794

so that jobs can be submitted and queued when requesting more licenses
than are configured. This is helpful when using licenses to control
access to file systems.

Bug 5630

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

so that a job's license string can be updated after a restart with licenses
removed or added.

Bug 5630

If a job requested multiple licenses and one license was removed then on
a reconfigure/restart all of the licenses -- including the valid ones
would be removed.

Bug 5630

when configured license counts are 0.

Bug 5630

Bug 5680

determine the time a node has been down.

This prevents from the wrongly set node as DOWN, if we decrease
SlurmdTimeout and agents queue is busy.

Bug 5811

it is reconfigured to determine changes in its background timers.

Bug 5811

reached in the cons_res plugin.

Bug 5328

structures after unpack.

Bug 5821

Bug 5815

Also simplify this xstrfmtcat() into an xstrcat() call while here.

Bug 5828.

Leads to false-positive warnings about settings being too low when
they're actually fine. E.g., if innodb_buffer_pool_size is over 2GB
the warning would have been printed in error.

Bug 5661.

Bug 5594.

You should not make RPC calls within MailProg (or anything else
fork()+exec()'d from slurmctld), so use the environment variable that
is now exposed instead.

Bug 5443.

smail currently uses 'scontrol show config' to get this data; making RPC
calls within programs fork()+exec()'d from slurmctld is best avoided as
it amplifies RPC load on the controller.

Bug 5443.

Add info to man page as well.

Bug 4961.

Bug 5724

Bug 5708

Continuation of 06582da8 (17.11.9)

Poll was timing out too quickly due to an incorrect conversion of
MessageTimeout. Added a multiplier so timeout reflects the correct
millisecond value.

Bug 5553

Since sacct exits the program when it finds unknown parameters, the
debug2() statements should be escalated to error() statements.
The first debug2() even had "Error: " in it.
Get rid of extra newline.

Bug 5421

Bug 5743.

Co-authored-by: Dominik Bartkiewicz <bart@schedmd.com>

The format for this field is "<inst>:<svc_id>:<comp_id>"; if inst or svc_id
are unset then they're blank, and the parsing code had assumed they would
always be blank as the systems slurmsmwd were developed on never used them.

Bug 5411.

Prevents job priorities from being lowered to 0 or from underflowing.

Bug 5165

Bug 5231

Change slurmctld/controller.c's purge_thread_lock from static to
global extern, so the mutex can be used in slurmctld/job_mgr.c

slurm_cond_signal() needs to always be wrapped in the same mutex as
slurm_cond_[timed]wait(), or else there is a possibility that
slurm_cond_signal() will trigger before slurm_cond_[timed]wait()
is even listening, most likely causing a deadlock.

initialize old fairshare association pointer correctly.

Bug 5744

Bug 5458