Bug 4405

Follow on to a37f8a1d for slurm.spec.

Avoid duplicating this logic in the Source and setup definitions.

Bug 4421.

%systemd_preun_with_restart was a typo, replace with
correct %systemd_postun_with_restart macro.

Bug 4421.

This reverts commit 29ac92fc.

Restore the expected and long-standing practice of having the main
slurm package own /etc/slurm.

Bug 4421.

Bug 4433.

since the ABI could change from one version to the other.

Bug 4435

Bug 4431

This primarily happens regularly if

PrologFlags=contain

Where the output has a nodename (and usually a number in it)

Normal output is

salloc: Granted job allocation 31159
1

but if the allocation is delayed you get output like

salloc: Granted job allocation 31161
salloc: Waiting for resource configuration
salloc: Nodes snowflake4 are ready for job
1
salloc: Relinquishing job allocation 31161

Where the test would key on the number in snowflake4 and fail.

This change makes it so the number has to be the first thing seen.

With the change to dynamic linking for libslurmfull, all
packages must match not just the API version but the exact
release. Set this for all sub-packages in accordance with the
Fedora packaging guidelines.

Bug 4426.

bug 4356

Continue to build with OpenSSL by default, which can be disabled
with --without openssl.

No behavioral change.

Bug 4428.

_top_priority() is used to determine at job allocation time (called from
job_allocate()) if any other job has a higher priority than the specified
job being allocated. If so, it sets the job state_reason to WAIT_PRIORITY.
This patch skips the priority comparison if the two jobs being compared
are pack job components from the same heterogeneous job.

Bug 4400.

Global replace operation made some undesired changes

bug 4247

in _unpack_node_info_members().

Bug 4389

Add  special case for SELECT_PLUGIN_CRAY_CONS_RES, SELECT_PLUGIN_CRAY_LINEAR
to select_get_plugin_id_pos()

Bug 4389

If a pending federated job exists on clusters 2 and 3 and squeue is run
from cluster 1 then the active siblings can come and go depending on
which cluster returns the job info first and depending if that cluster
is the origin cluster or not. The origin cluster only knows where the
active siblings are.

from status commands.

Bug 4341

Add in strong_alias calls so these functions will appear in
libslurm and not just libslurmfull. Otherwise test7.3 can fail
due to a missing symbol if a gpu gres is allocated to the job.

Bug 4415.

Bug 4404.

This setting causes /unix to be omitted from the xauth string.
Simplify the regex to handle this by adding / to the earlier match and
dropping the /unix pattern.

Bug 4417.

bug 4400

bug 4256

Bug 4379

Can cause slurmstepd to crash, as rlimit_name was pointing to
part of the free'd env_name variable.

Bug 4409.

This patch has fixed the problem to me. We are going to do some more
verification later today and update.
But I would appreciate if somebody else can test it as well.

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

There was a list of pending pack job records under consideration
  for scheduling by the backfill plugin that was not being cleared
  between interations of the backfill scheduler resulting in various
  scheduling anomalies.
bug 4371, 4400

An abort was triggered here due to a pack job start failure.

code dealing with how we need to keep track of it.

Bug 4405

For heterogeneous job steps, the srun --open-mode option default value will
    be set to "append".

Previous logic would fail if more than 2 pack groups implicitly
  specified in srun command.

Previous logic would delay the initiation of pack jobs until all
  components were submitted. The new logic will defer pack job
  scheduling based upon a new "pack_job_offset" field in the
  job submit request and NOT set a begin_time in the future.
  This will eliminate the pack job scheduling reason value
  of "BEGIN_TIME".
bugs 4369, 4400

Fix for bug introduced in commit 9e0b976a
bug 4400