The other responses have so far talked about hardware setup, so I’m not going to do that. Instead I’m looking at your software setup: VMs can be comparatively power inefficient compared to containers, specially for always-on services that idle often.
Does registry still have that problem of making it practically impossible to do garbage collection on old images?
Yeah, you’d have a LoadBalancer service for Traefik which gets assigned a VIP outside the cluster.
virtual IP addresses
Yeah, metallb.
The container is reproducible. Container configuration is in version control. That leaves you with the volumes mounted into the container, which you back up like any other disk.
It’s not that Seagate improved (which it may have), it’s more that WD has noticeably declined. It’s not a race to the bottom (yet), but there’s effectively no competition any more, so they aren’t incentivised to improve quality.
Figure out the uid/gid (numeric) for the user in lxc, then change the data permissions to those.
Use -m and limit the build job’s memory so it doesn’t kill the docker daemon.
Or they work in a regulated industry that requires pseudo-airgapped machines for remote users, e.g. the machine actually interacting with the systems needs to be within the controlled boundary but the company has a presence in multiple locations, so the solution is to have a Citrix server that the users remote into. But because the SSP also has access control requirements at every stage that take a long time to get updated to newest industry standards, the user still needs to have passwords rotated, MFA, and all that kaboodle.
The one thing snap does that flatpak doesn’t is provide CLI applications. But then nix also does that, so snap can go pound salt.
I used a Fractal Design case for a home server in the past. Pretty happy with them.
Renders in Summit, but I turned that off because any post title starting with
#is rendered as h1.