Hacker News Reader

Comment by amelius | Hacker News Reader

Is this how embedded folks make sure that a device starts with exactly the same installation every time a machine is booted?

I wonder why embedded products like Nvidia Jetson do not come with an immutable Linux (and instead are based on Ubuntu which updates itself on every opportunity via apt and snap and whatnot).

View on HN

There are lots of companies using NixOS for this, BalenaOS (Yocto + Docker), or building their own bespoke tooling on top of a minimal Linux setup.

Although many places start with Ubuntu or Debian in my experience it’s common to invest a lot of time and energy in getting out of that unmanaged setup once the company scales.

The hardware usually comes with vendor-specific libraries (e.g. cuda in the case of nvidia) which are based on a specific version of libc, so then you will have to build your entire alternative OS around that version also.

Which is… never trivial. I’d say 25-50% of my career so far has been repeatedly “fixing” clunky deployments of ROS, OpenCV, L4T, CUDA, cudnn, libc, etc. in Docker and Nix. Fun stuff!

It’s common for hardware vendors to provide a working system for demonstration purposes so you can evaluate the hardware without having to learn an immutable OS toolkit. Then when you pick hardware you also do the bring up work to get the kernel compiling from source and integrated with your userspace of choice. At that point you’ll switch to an immutable system.

Hardware vendors in this space can’t be trusted, so you need to make sure the board is actually fit for purpose. Outside of the hobbyist space you have to be really careful. There are often business objectives that rely on the board working a certain way.

This is nice in theory but the amount of vendor-specific libraries can be quite large (e.g. Nvidia's CUDA, libcudnn etc.), which you then have to get working on your new OS.

Reply on news.ycombinator.com