So I’ve molecule’d (via rootless podman) my draft Ansible playbook. It’s of course far from perfect yet. But I should have done it before.
Before diving a bit more into that subject, I’ve bought a second-hand (or third or whatever) AMD FX-8370E with 2×8GB DDR3 DIMM (unknown technical details at this point, I’ll receive it in 10 days or so) to pimp a bit my desktop PC (currently running a Phenom II x6). So I’ve unburied my previous dormant mobo (Asrock N68C-S UCC), which will inherit the Phenom instead of the Athlon X2 6000+ I’ve previously used, along with 2×2GB DDR2-800 DIMMS. Hopefully, that mobo is one of the seldom AM2 and AM3 CPU compatible, and it can run DDR3 on 2 slots with the Phenom II previously cited. I now need to find a cheap tower/PSU/graphic card to complete it even though a gbps network card may be another welcome addition too (onboard NIC is 100mbps max…). Anyway, let’s go back to the molecule-ization (?) of the Ansible playbook.
I’ve had to duplicate the playbook.yml to « shape » it into a task file or it wouldn’t run. Biggest joke of the day is that I want something which is SystemD aware (I need to run at least PostgreSQL and MAAS), and that’s not the case of containers (well, I’m not done with researches about SystemD and containers, but it looks like that won’t do the trick, I’m not fond of spawning multiples containers, that would diverge a bit too much from the final state). I’ll get to spawn a VM with molecule instead.
All in all, things are progressing. I’ll commit things to codeberg once I get:
- Molecule spawning a VM.
- Why relaunching MAAS install fails (an obscure « service maas-rackd is masked » making maas-region-controller installation fail)
- Firewall rules are done.
After that, I’ll commit more frequently, I understand you (if anyone reads this) can find things frustating. We’ll be able to provision and deploy the cluster (given, I bet, quite a bunch of debugging Dash support).