Experimenting with FreeBSD and comments on the future of Invidious

published on 2024-09-18 by hyperreal

Experimenting with FreeBSD

I’ve always admired the BSD family of operating systems even though I mostly use Linux. I currently run a Debian Linux server that hosts a bunch of alternative services like PrivateBin, Invidious, RedLib, etc. Most of these require the use of Docker and/or systemd. Docker doesn’t seem to be a hard requirement, though, because inside the Dockerfile there has got to be a Unixy way to setup and run the services. So I’m planning on spinning up a temporary FreeBSD VPS on Vultr and experimenting with getting those services running without Docker. If all goes well, I will convert my main Debian server to FreeBSD. One thing I would miss, however, is FirewallD. BSD’s PF (Packet Filter) firewall doesn’t have a command-line interface to interact with firewall policies. As far as I know everything goes into a /etc/pf.conf configuration file. I’d have to look into PF configuration options that are effectively the same as the ones I use with FirewallD.

Comments on the future of the Invidious project

Re: Invidious, it’s currently broken on my instance because of Google blocking the use of the API to by-pass the ads. From what I understand, it’s a bit of an arms race between the Invidious developers circumventing the Google blocks and Google catching on and imposing restrictions. I’m not sure about the future of the project. Hosting an Invidious instance, and keeping in line with the requirements to remain an official instance recommended by the project in their instances list, heavily relies on Docker workflows, so it might not be feasible for me to get it running on FreeBSD and continue to satisfy those requirements. I don’t think it’s worth running an instance until a more permanent solution to the Google blocking problem is found, anyway.

See this issue on the Invidious GitHub repo for more information on what’s going on with it. Every time there is a fix, a new problem pops up that breaks Invidious instances. This comment on the GitHub issue was the last straw for me. Previous fixes require using and depending on a bunch of hacky tools from the developers that aren’t guaranteed to be long-term solutions. I very much appreciate the time the devs have put into the project. I really, really hate Google’s unrelenting mission to break the fixes in order to maintain their monopoly on video content. YouTube is just too popular to fail at this point, and Google will never let it fail. The only way forward that I can see from here is just to promote PeerTube in the fediverse or other FOSS video hosting alternatives.

Links