Turns out if the extern step is created here and the job was killed long
before hand the step is made erroneously and can cause an assert just lines
later.

Bug 3898

specify an alternative --ntasks-per-*

set correctly.

Bug 3858

Changes the alpsc_configure_nic() call to set the exclusive flag,
and 100 for both the cpu and memory scaling values.

Should only be used with exclusive jobs without concurrent steps
running on a node, otherwise oversubscription of the GNI resources
can occur leading to performance issues.

Bug 3713.

What this does is populate the node_hash_table as nodes are being read in
instead of after the node_record_table_ptr has been fully populated.

This speeds up a start of a slurmd with a system of 10000 nodes from
> 1 minute to less than a second.

In 17.11 we will remove the linear xstrcmp check as it should no longer be
needed.

Bug 3885

code change.

Bug 3885

batch job takes longer than it takes to finish.

Bug 3833

time.

Bug 3833

An array was only being partially cleared due to bad logic
bug 3876

Bug 3874.

Improve selection of jobs to preempt when there are multiple partitions
    with jobs subject to preemption.
bug 3824

Prevent segfault from pointer dereference to the QOS that is
being deleted.

Fix to commit 3e8aa451.

Fix regression from commit c05dcb8a (bug 1923) that doesn't take
into consideration a blank char * as a valid option.

This fixes the scenario like

sacctmgr list associations user=''

which would only print account associations.
Bug 3862

a good return code.

This also fixes the situation where the step was ending but not yet ended
so it sends the KILL_TASK_FAILED error instead of JOB_NOTRUNNING.

Also it removes the abort in favor of exit which it should had been anyways.

Bug 3758

list_for_each)

Inadvertently set to one when requested.

Bug 3855.

Bug 3857.

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.

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.

Bug 3840.

node_featurs/knl_cray plugin: Don't clear configured GRES from non-KNL node.
bug 3768