Bug 4323

Bug 4408.

Bug 4306.

to start.

Bug 4434

Bug 4406

Bug 4405

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

Bug 4435

Bug 4431

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.

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

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

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

The regex in x11_set_xauth() did not match FQDNs because it was missing
a dot.

Bug 4398.

Add SchedulerParameters=whole_pack configuration parameter. If set, then
    hold, release and cancel operations on any component of a heterogeneous job
    will be applied to all components.
bug 4374

Bug 4393.

bug 4366

Correct printing error type based upon errno rather than returned rc.

If PrologSlurmctld fails for pack job leader then kill all
    components of the job.
bug 4379

Add SLURM_PACK_JOB_NODELIST to PrologSlurmctld and EpilogSlurmctld
    environment.
bug 4379

bug 4370

Prevent scheduling deadlock with multiple components of heterogeneous job
    in different partitions (i.e. one heterogeneous job component is higher
    priority in one partition and another component is lower priority in a
    different partition).
bug 4370

From within slurm_job_submit():
	job_desc.pack_job_offset

From within slurm_job_modify():
	job_rec.pack_job_id
	job_rec.pack_job_id_set
	job_rec.pack_job_offset

Bug 4372.

bug 4368

Add SLURM_PACK_JOB_ID and SLURM_PACK_JOB_OFFSET to PrologSlurmctld and
  EpilogSlurmctld environment
bug 4379

Avoid srun abort trying to run on heterogeneous job component that has
    ended.
bug 4366

In a prior incarnation of the patch that introduced it,
it was MaxQueryTimeLimit, and that was not updated with
the code base when changed.

Bug 4365.

This now matches the sinfo documentation.

Bug 4306.

launch/slurm plugin - Avoid using global variable for heterogeneous job
    steps, which could corrupt memory.
bug 4333

Ancient versions of OpenMPI and their derivatives (i.e. Cray MPI) are
dependent upon communication ports being assigned to them by Slurm. Such MPI
jobs will experience step launch failure if any component of a
heterogeneous job step is unable to acquire the allocated ports.
Non-heterogeneous job steps will retry step launch using a new set of
communication ports (no change in Slurm behavior).

NOTE: Correcting this would necessitate assigning the same set of ports
to all components of the heterogeneous job (not possible today) plus changes to
srun in order to better synchronize the step startup and error handling.

Same logic as done in commit fb296c70 done for energy.

Bug 4336