Increase the sleep in job to 5s was required to make test reliable. It
doesn't result in longer execution since we're waiting for job to get
into RUNNING state and then we're canceling it. Short sleep can end-up
with overlook of RUNNING state and false negative result - 'Job is DONE
but expected RUNNING'.

Bug 7624.

Previously there were no log_error before setting exit_code to 1
due scontrol error.
Now user can identify the reason of final FAILURE result.

Bug 7624.

Bug 7987

Co-authored-by: Broderick Gardner <broderick@schedmd.com>
Signed-off-by: Broderick Gardner <broderick@schedmd.com>

This avoids possible overlaping with other jobs.

Bug 7661.

Bug 7712

Bug 7723

Bug 7664

Bug 7149.

The get_default_acct was truncating the account name for long account
names. This commit uses -P/--parsable2 to avoid it.

Bug 7369.

Fixes a compilation issue with the job_container/cncu plugin.

Bug 7398

Bug 6772, 6977.

Update slurm.spec and slurm.spec-legacy as well.

Use safe_unpackstr_xmalloc() instead of unpackstr_ptr to ensure all
strings are escaped correctly in SlurmDBD.

This requires the introduction of the _free_local_*() functions to free
the structure members, whereas they were previously just pointers into
the single managed buffer which was cleaned up elsewhere.

Remove the unsafe unpackstr_ptr macro definition as well while here.

CVE-2019-12838.

Flags are stored in a smallint, which can only hold the first 16 bits
worth out of 32 bits of flags currently in use.

MySQL's overflow rules will treat any value > 0xffff as 0xffff, rather than
dropping the higher-order bits (flags), which means the stored value not
only loses the higher-order bits but corrupts the lower-order as well.

The 19.05 release extends the column to bigint (64 bit).

Bug 6969.

For heterogeneous jobs, do not count the each component against the QOS or
association job limit multiple times.

bug 7190

Bug 6847

Catch jobs with empty pack_job_id_sets.
Corruption bug fixed by previous commit.

Bug 6847

Rejected hetjobs were not correctly dump because pack_job_id_set was
not properly set in _slurm_rpc_allocate_pack() when submision was
rejected.

New helper function _create_pack_job_id_set() to avoid duplicated
code.

Bug 6847

Fix for test 21.39. If the job didn't start running before the
account/qos was removed, then the job would never start running, and the
test would fail.

Bug 6942.

Previous logic could hang indefinitely if srun did not return
It also performed unneeded test setup before confirming that the
system size was adequate to execute the test.

bug 7050

Remove excess tabs to properly align code
Make comment formats consistent

No functional changes.

Bug 7050.

Bug 7050.

Bug 7050.

Handles cases where last line of slurm.conf is not empty.

Bug 7050.

One part of this test can fail in a heterogeneous system. Do not run the
part of the test that would fail if the system is heterogeneous.

Bug 7051.

backport test7.20 changes from master to v18.08.

bug 7050

select/cons_res is recommended.

Bug 7138

Bug 6987.

Bug 6466.

Bug 7101.

The name variable hasn't been set yet, so this is always NULL. Print the
uid/gid instead. While here, treat uid/gid as uint32_t, and use strtoul()
rather than atoi() to avoid issues with high-number uid/gid values.

Fixes GCC 9 warning.

Bug 7101.

Bug 7101.

Use RETRY_DELAY per to mirror job complete delay but without a max retry
count for the time being.

Bug 6970.

Signed-off-by: Brian Christiansen <brian@schedmd.com>

Bug 6964

The reason was being set after the message was sent to the db. Also
clear the draing and reboot states before the message is sent so that
the event state will show DOWN.

Bug 6964

Bug 6964