Previously it was taking the MIN, without respecting the order.

Also add a note to the resource_limits.html page to clarify the
exception for Max[Wall|Time] and/or [Max|Min]Nodes limits, where
the default is that the Partition is the king with regards of
precedence, unless the respective job's QOS flags
Partition[Min|Max|Time]Limit are set.

Bug 4681.

This prevents a hard-to-diagnose issue where slurmstepd may fail
to start due to a missing library. This now ensures slurmd will
fail, and keep the node down until the library issue can be fixed.

Bug 4645, 4644.

fatal() calls exit(1) which precludes getting a backtrace.
That's fine on configuration issues and other types of problem,
but for hitting "impossible" edge cases getting a core dump may
be the only way to isolate the issue.

Adding to 17.11 so we can easily provide diagnostic patches without
needing users to back-port this implementation. Further use will
come in 18.08.

Bug 4599.

This reverts commit 18b35709.

This reverts commit 6a91845b.

Wait for the prolog to complete, and then clean up after it.
Otherwise orphaned "sleep" executables will be created by the
still-launching step_extern.

Bug 4718.

Preserve node features when slurmctld daemons reconfigured including
active and available KNL features.
bug 4734

Bug 4630.

Bug 4722

keyvalue_initialized is reset on 'scontrol reconfigure' and other
cases, which can lead to additional atfork handlers being registered.

These can eventually lead to a segfault if an excessive number of
handlers have been re-registered. Set a separate boolean to protect
against this. Clear that boolean as part of the atfork handler.

Bug 4628.

UsePss was correct, but UsePSS and usepss would be silently ignored,
leading to confusion as to whether the option was working or not.

Treat all JobAcctGatherParams as case-insensitive to avoid confusion.

Bug 4637.

This reverts commit 516b0d59.

With the fixing of the NEWS file.
We want to keep the idea of only checking one federation.

The argument is redundant and possibly confusing for pack jobs

Bug 4548

This reverts commit fb73b8a4.

# Conflicts:
#	NEWS

message before purging the job record to get the uid of the revoked job.

Bug 4502

This reverts commit fc0c3e6c.

message before purging the job record to get the uid of the revoked job.

Bug 4502

The value is redundant and possibly misleading for pack jobs

The information is redundant, and possibly inaccurate for
  heterogeneous jobs

Bug 4651

Bug 4548

Bug 4634

job container where if the step was canceled would also cancel the stepd
erroneously.

Bug 4634

one already exists owned by a different user will be logged and the job
held.

Bug 4614

Bug 4681

In preparation for release of v18.08.0-pre1 later this week.

from when last started.

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

invalid reads.

Bug 4664

if LaunchParameter test_exec is set.

Bug 4439

rights by a secondary group id.

Bug 4439

before the primary takes control.

The primary will wait until the backup has replied with it all the way
done shutting down before it takes over.  Before it would just wait 2
seconds which probably was ok in most situations.  Now we don't return
until it is completely out.

This also ups the CONTROL_TIMEOUT from 10 -> 30 seconds.

Bug 4446