Describe cgroup-based core specialization support.

Conflicts:
	src/slurmctld/job_mgr.c

If the backup slurmctld assumes primary status, then do NOT purge any
job state files (batch script and environment files) but if any attempt
is made to re-use them consider this a fatal error. It may indicate that
multiple primary slurmctld daemons are active (e.g. both backup and primary
are functioning as primary and there is a split brain problem).

Replace printing of job_id using %d with %u

Conflicts:
	src/slurmctld/slurmctld.h

Test time when job_state file was written to detect multiple primary
slurmctld daemons (e.g. both backup and primary are functioning as
primary and there is a split brain problem).

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

when specified escaped.

original add and  mod functions.

like the switch plugin functions. When a batch job runs a core
specialization job on the same node, it was failing because both
tried to use specialized cores at the same time.

Conflicts:
	slurm/slurm_errno.h
	src/common/slurm_errno.c

A configuration change trigger event occurs when a node state
changes (e.g. Up, Down, Drain, etc.)

This reverts commit f3a66dd8.

This reverts commit 421185f4.

plugin, which call corresponding alpscomm functions. This enables job
steps to give up some of their network resources on suspend, increasing
the amount of resources which can be given to each job.

Attempt to create duplicate event trigger now generates ESLURM_TRIGGER_DUP
("Duplicate event trigger").

Modify strigger to accept arguments to the program to execute when an
event trigger occurs.

Added strigger option of -N, --noheader to not print the header when
displaying a list of triggers.

batch jobs have cpus_per_task set to zero, which resulted in an
error of "task/cgroup: task[0] unable to set taskset '0x0'"

Recover a list of running jobs when slurmd restarts. This job
list is used to determine when a job suspend can take place.
This patch also adds job suspend suspend/retry logic since
a job suspended immediately after launch can (briefly) return
an error that is it not ready yet.

If the job to be suspended can not be found, then proceed with the
suspend RPC after 3 seconds, Previous logic would hang indefinitely.
The job would not be found if it was launched, then the slurmd restarted.
The record keeping for launched jobs after slurmd restarts is fixed
in a separate patch.

This reverts commit bf83bcf7.

This reverts commit 4294c561.

requeue, requeuehold and release operations.