Scrub status "unknown" - Leap 15.6 OS base #2872

[phillxnet]

Adopt packaging library as a direct dependency to enable more sophisticated btrfs-progs version assessment. Previously already an indirect dependency of gunicorn. Moving to explicit dependency to avoid issues re gunicorn's deprecation in the future.

Includes

• Additional test data pertaining to Leap 15.6 and newer, re btrfs-progs version.
• Additional test data re exception handling (fail through) re non PEP 440 compliant 'btrfs-progs version' output.

Fixes #2872

[phillxnet]

Testing

On an issue-reproducer system (dev machine using Leap 15.6 X86_64) we have one before (see: #2872 (comment)) and one after scrub status reports:

And we have the following imported ROOT pool in "running" scrub status:

After a little time, and Web-UI refresh, we see this same ROOT pool scrub as having finnished:

[FroggyFlox]

Thanks @phillxnet for this proper fix!
I can confirm it fixes the issue in Leap 15.6 on my end too; Leap 15.5 is still working as intended as well.
All tests also pass on my Leap 15.6 VM.

[phillxnet]

@FroggyFlox Thanks for the review & 15.5 check: much appreciated. I'll get this merged directly.

I know we have the question of need, for this legacy btrfs-progs support, as per #2872 (comment): i.e. do we really need to support pre Leap 15.4 btrfs-progs versions. But we are in very late testing here so minimal changes to address indicated issue. We also have the potential to migrate scrub to Huey (or a more Django native Huey replacement) but our consensus appears to be that this is better tackled in the next or next-plus-one testing phase.

[phillxnet]

An rpm was test built using this branch (Leap 15.6 X86_64 host) and looks to behave as-per the testing report so far. We also have the new dependency here which now shows up as a new top-level one re:

cd /opt/rockstor/
poetry show --tree
...
gunicorn 22.0.0 WSGI HTTP Server for UNIX
└── packaging *
...
packaging 24.1 Core utilities for Python packages
psutil 5.9.4 Cross-platform lib for process and system monitoring in Python.
...

Where-as previously we only saw the packaging dependency as secondary: i.e. via gunicorn as in the above.