This adds the ability for a step to explicitly clear a job's gres counter.
For example, the job requests --gpus-per-task=2 and the step requests
--gpus-per-task=0.

change some comments and log messages.
no change in functionality

failing to propagate gpus-per-task for job step

Move reused GRES parsing logic to a shared function.

(Corrected author; was previously mis-committed as 433257c6.)

Bug 5743.

This reverts commit 433257c6.

This just tests the gpus-per-node option. Additional tests are
needed for gpus-per-task, etc. as well as the logic to support
those options for job steps.

If a job with running steps is resized, that will result in the job's
core_bitmap size changing, but not that of the step, resulting in an
error message. This adds a job resized flag and prints a more meaningful
message when appropriate.

change to bit_ffs/bit_fls rather than full scan, improves performance

check for TaskPlugin != none at start
Print all hostnames associated with launched job (to diagnose problems)

Coverity CID 187779

Previous logic could abort in some cases

The last merge from v18.08 (c6f696db) included changes that
are incompatible with cgroup configuration read function changes
that were made in master.

sinfo output in this test is grouping all nodes together in one
output line (e.g. "nid[000001-000099'), while the data used to
have one node per line. Adding the "-N" option to the test returns
the output to one node per line, as required for the test to work.

Bug 5750

when the node_bitmap scan was changed from checking all nodes
to just the nodes within the range set, the node_ptr became
wrong, resulting in bad job counts on nodes and nodes marked
as allocated when there were no jobs active (bug only in master).

No change in logic

Move reused GRES parsing logic to a shared function
bug 5743

This shouldn't happen in practice.
bug 5749

Test1.113 was sometimes failing due to timing issues related to
the Cray Node Health Check running. Add small sleep to (mostly)
eliminate the problem.

The option was previously ignored with the cons_res plugin plus
topology/tree. With cons_tres, handle the option by ignoring the
topology and using the default node selection logic.

add logic to provide optimized resource allocation with respect
to topology/tree configuration. Mostly working
Fix max-node enforcement

Introduced in commit 3e727cc4
Coverity CID 188211

Problem introduced in commit a044d01b
Coverity CID 188212

They were being referenced prior to initialization

This better handles job submit failures and reduces the time spend
waiting for output files by overlapping the wait for all files.

without this the total GRES counter can keep getting incremented,
the _get_job_cnt() calculation can overflow and extra GRES be
allocated to the job

# Conflicts:
#	src/common/slurmdbd_defs.c

Bug 5544

Bug 5544

Bug 5613

This reverts debugging accidentally committed in 48ec9695

There are a few anomalies with respect to front-end mode, which
test test changes address