WIP: Feat/sms pulp

[sjpb]

• Seeing what speedup we get from a local pulp mirror on SMS. Deployed as per docs.
• Only RL9 synced.
• Would need pipeline to sync pulp on snapshot updates

[sjpb]

Image build: https://github.com/stackhpc/ansible-slurm-appliance/actions/runs/22186686411

• Above RL9 build attempt 1 1h3m, 1h 3m, 1h 0m, 1h 0m
• Previous successful RL9 builds: 39m10s, 55m22s, 1h32m (was timestamp bump, why does that matter), 1h11m, 1h14m, 1h24m, ...

[sjpb]

Changed pulp server from 2vpcu (general.v1.small) to 8vcpu (general.v1.large)

[sjpb]

Image build: https://github.com/stackhpc/ansible-slurm-appliance/actions/runs/22436458273
attempts 1, 2 failed - needs a sync due to updating from main
DONE: sync for RL9.
attempt 3: 53m 1s 5: 48m 42s, 6: 1h22m

edit: didn't see load averages > ~1.4 on the pulp server during image build, so don't think pulp CPU is the issue here

on the 2cpu build vm near the end, /proc/loadavg showed : 1.68 1.27 1.41 7/449 50740

[sjpb]

checked build uses ssd voluem, OK

Image build w/ 4x CPUs: https://github.com/stackhpc/ansible-slurm-appliance/actions/runs/22447693337

attempt 1: 58m