Bug 3857.

File deletion can be slow, especially when StateSaveLocation in on
NFS or other network filesystems. Since purge_old_job() holds all
the slurmctld write locks, this is especially performance sensitive.

Moving this to an independent thread lets the slower filesystem
cleanup happen without owning these locks. purge_old_job() then
results in the purged job ids being queued in the purge_list.

A race with the job id potentially wrapping around again is already
prevented by _dup_job_file_test() in get_next_job_id().

Bug 3763.

Bug 3840.

node_featurs/knl_cray plugin: Don't clear configured GRES from non-KNL node.
bug 3768

the SchedulerParameters=reduce_completing_frag option.

NOTE: reduce_completing_frag on or off only works with CompletingWait set to
something.

Bug 3756

For jobs submited to multiple partitions, report the job's earliest start
    time for any partition.
bug 3754

fragmentation.

Bug 3756

Two jobs completing simultaneously leads to make_node_idle()
returning before it has a chance to decrement node_ptr->owner_job_cnt,
which can result in the node being "owned" by that user even
through no jobs are running on it.

Move the decrement block to the end at a fini label, and make sure
all return paths pass through it. While moving that add a guard
against node_ptr->owner_job_cnt underflowing.

Bug 3771.

If a job is considered on a partition with ExclusiveUser=YES
then it would be marked as if it was submitted with the
--exclusive flag, which would lead to delays launching it
on ExclusiveUser=NO partitions, and cause lower-than-expected
cluster usage.

As a side effect, the job_ptr->part_ptr->flags need to be
tested wherever WHOLE_NODE_USER is considered, instead of
just job_ptr->details->whole_node.

Bug 3771.

Wrong author attributed by mistake.

This reverts commit 9128476a.

Wrong author attributed by mistake.

This reverts commit a02d04f1.

leaving the node owned.

Two jobs completing simultaneously leads to make_node_idle()
returning before it has a chance to decrement node_ptr->owner_job_cnt,
which can result in the node being "owned" by that user even
through no jobs are running on it.

Move the decrement block to the end at a fini label, and make sure
all return paths pass through it. While moving that add a guard
against node_ptr->owner_job_cnt underflowing.

Bug 3771.

WHOLE_NODE_USER.

If a job is considered on a partition with ExclusiveUser=YES
then it would be marked as if it was submitted with the
--exclusive flag, which would lead to delays launching it
on ExclusiveUser=NO partitions, and cause lower-than-expected
cluster usage.

As a side effect, the job_ptr->part_ptr->flags need to be
tested wherever WHOLE_NODE_USER is considered, instead of
just job_ptr->details->whole_node.

Bug 3771.

_setup_assoc_cond_limits was using the table 'prefix' passed by argument
in the where clause to select the where clause prefix.deleted=something.

It turns out that _setup_assoc_cond_limits is called by these functions:
as_mysql_modify_assocs
as_mysql_remove_assocs
as_mysql_get_assocs
as_mysql_acct_no_users

which set the prefix to 't2' before the call if a QOS is provided or if
WithSubAccounts is provided. The 't2' prefix is fine for other where
conditions in that case, but for choosing the deleted we need the t1
which is the table we're selecting the records off.

Bug 3835

Bug 3821

'scontrol update reservationname=foo duration=unlimited' sets INFINITE
as the duration, which needs to be translated to a year as is done
elsewhere. Otherwise it'll convert to 49710 days, which is definitely
wrong.

Bug 3836.

Bug 3443.

MAX() will re-evaluate the higher value argument; if this is a function
is may be called twice over, leading to unintended side effects or a
crash.

Bug 3443.

This also fixes the fed_mgr on the backup slurmctld to start backup
correctly if the backup takes control more than once.

Bug 3827

Previously, incorrect partitions and duplicated partition names
could be shown.

The array needs to be incremented by two, not one, as each
element is a start-end pair.

Bug 3793.

when configured with memory-leak-debug

Bug 2131

Bug 3776

KNL modes if not KNL system.

Bug 3825

This reverts commit c2380520.

KNL modes if not KNL system.

Bug 3825

reconfigured while node boot in progress.

Bug 3679

bug 3822

Bug 3708

In 17.11(or other future version) we should move a lot of this common
code into a new lib.  The reason I didn't put these common changes
into common/xlua.c was because then I would have to link common to
liblua which I really didn't want to do.

dependencies.

Bug 3708

Bug 3772

are ready.  This will be NULL is a job was revoked while beginning.

if there was a failure after the initial resource allocation. Does not
release previously allocated resources.

Bug 3783

This is the initial patch that adds the stubs for the logic.  Outside of
that this patch really does nothing.

configured.

fails to enter the desired NUMA and/or MCDRAM mode then drain the node and
requeue the job.

Bug 3785