The temporary OpenBSD cloud server hints at a solution. When I go away, I can fire up (ho ho) cloud servers, and destroy them when I return. If they explode, they explode in the cloud, not chez nous. If I do this properly, I’ll have a number of temporary servers to set up and take down, as and when. Doing all that configuration by hand is bonkers, especially if they process is likely to happen more than once.

The electricity supply to the flat is imperfect, mostly because a substation nearby recently blew up (see, my fear of electrical explosions isn’t entirely bonkers). That’ll get sorted, but, for the time being, the flat is occasionally blacked out for a few hours. That is a reason to keep some servers in reserve for immediate operation, probably on Azure, given Microsoft only charge storage costs for servers that aren’t running.

I have configuration issues with my systems now. They’re all somewhat similar, but somewhat different in configuration and presumptions, because I set them up at different times for different reasons. Now, these are home systems, so that’s not actually a big deal, but it is untidy. I would like to get them under control.

All these things show why I need an orchestrator. I decided to use ansible because it works on a good number of operating systems, including OpenBSD, and because I’ve used it before. It works, and most of its foibles can be avoided, so it’ll do.