Native install documentation?

[lordkitsuna]

Hello, looking at possibly using this in our infra. however looking over the documentation i am only seeing instructions on using Docker or Kubernetes. I assume since this project is rust i can just use cargo to compile it and see reference to that in some issues. However i want to just verify that would be enough to get a fully functional install and i could use the same documentation for the rest of the configuration. Our infra uses unprivileged and filtered LXC for its containers and we would really rather not do nested docker (especially since docker networking would not work in unprivileged containers)

[sebadob]

(...) and we would really rather not do nested docker

Right, I would never do that as well.

Instead of docker, you can of course use any compatible runtime like podman or even apple container. However, yes of course you can run it natively without issues. Running it bare metal is even the fastest and most efficient approach. The only reason it's conatainerized by default is that I usually run everything in Kubernetes. I like to treat all my compute as ephemeral and containers make it easy. That's the only reason. There is no native installer (yet). I would maybe create that when Rauthy reaches v1.0.0.

Compiling from source is pretty easy as long as you have some dependencies installed on the OS. All of that is covered in the CONTRIBUTING.md. When you just want to compile from source you technically would not even need npm, as you can use the archived frontend files that are mentioned in the Free BSD section.

Another option would be to pull an existing image for your arch and extract the binary out of it. The outcome is the same, and this may be your preferred method because it might be easier / quicker for you. The images do only exist for Linux x86_64 and arm64. If you are running on another arch / OS, the only option is to compile from source.

CAUTION

If you are compiling from source, and you want to use that version in production, NEVER compile from the main branch! Always check out the specific version tag. The main branch may contain unstable database migrations, which have not been finalized yet. If these are applieid to a DB and they need any changes before the next release, you will need to recreate the whole DB each time, or do a fully manual rollback.

[lordkitsuna]

Perfect! thanks for the detailed answer, we are using linux on x86_64 so just extracting from an image will probably be the simplest way forward for us. We used to use docker a lot but i got tired of how it buchers networking and wanted something more linux native and easier to hard lock down. so we ended up moving to linux containers (lxc) its been great, everything is pretty easy, other than so many projects being docker by default these days lol.