Crossing the threshold - Liminix#
Wed, 19 Oct 2022 21:20:32 +0000
I am restarting/rewriting NixWRT, which has seen no real development in, erm, about four years (my, how the time has flown) and is showing its age and showing my Nix inexperience.
līmen (genitive līminis) (neut.)
- threshold, doorstep, sill (bottom-most part of a doorway)
- threshold, entrance, doorway, approach; door
- house, home, abode, dwelling
- beginning, commencement
- end, termination
Thus: Liminix, which stands at the threshold of your home network. According to the commit history I've been playing around with it for about a month now (so, since shortly after I broke the family internet for most of a morning while trying to upgrade OpenWrt ), so although it still doesn't actually do anything useful yet perhaps it's time to break cover.
The objectives are quite similar to the NixWRT objectives in that I want to have congruent configuration management on the "infrastructure" devices that make up my home network, and those devices are typically underpowered for running full-blown NixOS. I do though have a shopping list of things I want to do better/differently:
- a writable filesystem so that software updates or reconfiguration (e.g. changing passwords) don't require taking the device offline to reflash it.
- more flexible service management with dependencies, to allow configurations such as "route through PPPoE if it is healthy, with fallback to LTE"
- a spec for valid configuration options (a la NixOS module options) to that we can detect errors at evaluation time instead of producing a bad image.
- a network-based mechanism for secrets management so that changes can be pushed from a central location to several Liminix devices at once
- send device metrics and logs to a monitoring/alerting/o11y infrastructure
So far: we're using s6-rc for services, which seems to be quite nice and well-put together but I haven't tried too hard to hurt yet. We're using the NixOS module system infra for declaring configuration option types and merging logic. We have significantly more in the way of automated testing than NixWRT had - admittedly not a high bar - and an entirely unrealised/untested idea of how we might do secrets. And the "we" there is, yes, editorial
We don't yet have: writable filesystem (ubifs?); anything o11y; more than one hardware device. And it's not yet at the point that I can dogfood it. Although technically it boots and runs on my spare GL-AR750, I haven't ported wifi across yet.
The primary repo is at https://gti.telent.net/dan/liminix because the older I get the more stubborn I become about free "if you're not paying for it you're the the product" services, but there's a mirror on Github for everyone who's not me. Because federated Gitea is not yet an available thing, and I don't want to throw up all the barriers to contribution.