Our local implementation was being used instead of glibc,
and does have one subtle difference - it double forks(),
whereas glibc single-forks, leading to some slight differences
in process control behavior. Revert this and take a different
approach.

This reverts commit 6f45a2bf.

which the backfill test window expands. This can be used on a system with
a modest number of running jobs (hundreds of jobs) to help prevent expected
start times of pending jobs to get pushed forward in time. On systems with
large numbers of running jobs, performance of the backfill scheduler will
suffer and fewer jobs will be evaluated.

Bug 3790

This reverts commit da414931.

which the backfill test window expands. This can be used on a system with
a modest number of running jobs (hundreds of jobs) to help prevent expected
start times of pending jobs to get pushed forward in time. On systems with
large numbers of running jobs, performance of the backfill scheduler will
suffer and fewer jobs will be evaluated.

Bug 3790

Bug 3671

Bug 3808

# Conflicts:
#	src/slurmctld/job_mgr.c

bug 3846

purge_files_list.

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.

Only called from _list_delete_job once the MinJobAge has
passed.

This will need to be handled differently. The timeout can
lead to the purge process falling further and further behind
on high throughput systems if the number of job scripts that
can be deleted within a second is lower than the job submission
and completion rate of the cluster, eventually leading to
the MaxJobCount limit being reached.

Bug 3763.

it works better on multi-slurmd installs.

Should be fed1,2,3 and not fed1,2,2

Bug 3839

Revert some of my b50f4661. Elaborate on tradeoffs, and point
to HTC page as well which is a better location for this info.

This is better discussed in the high_throughput.shtml doc.

Also, "Contrain" is misspelled adding to the confusion.

To submit sibling jobs to clusters that don't have the specified
features.

Bug 3859

Bug 3640

Instead of waiting for all jobs to clear out of the system, wait until
all jobs are completed. This helps so that you don't have to wait as
long for the cluster to be drain and or removed.

Clusters in the federation could be different rpc_versions so each
cluster needs to talk each other's language.

Routes request to origin cluster if it isn't running the job.

to give more flexibility.

if a federated job.

Move from slurm_send_recv_controller_rc_msg to
slurm_send_recv_controller_msg. This allows scontrol requeue to be
rerouted to the origin.