Anyone knows how to properly start a multiseat wayland with a desktop environment or window manager running?
I just need simple stuff such as profile initialization of the user and if it’s possible to just share the same discrete GPU across multiple seats?
The end result? I want to isolate my current user space from the gaming space where I can just connect using moonlight/sunshine. I want it all headless.
Lol I’ve just came on this sub to ask about this. Sunshine has a tutorial in their docs on how to do this on X11 but I’m using sway so I’m kinda fricced. If I start sunshine from ssh, it correctly uses the headless wayland display but it doesn’t have rights to access the encoder for some reason.
You can achieve similar (it’s gonna be slow though) results with wayvnc. You need export 3 variables in order to set it up according to their FAQ: https://github.com/any1/wayvnc/blob/master/FAQ.md
Edit: noticed your crosspost with answers: https://lemmy.world/post/12888914, thanks for creating this post.
Edit 2: WayVNC can be more usable if you use mosh instead of SSH. It makes VNC less laggy. The downside is that you have to start sway in normal SSH and then start wayvnc in the mosh session, which is kinda inconvenient.
Is this even possible in Wayland? i thought it was an Xorg thing
I am very interested in doing this, with xorg or Wayland, and an 7900xt.
Sunshine does have tutorial on how to do this for xorg in their docs. Haven’t tried it though, I’m too lazy to spend time setting up my configs for xorg when I already have a wayland setup.
I may have to give it a go. I am similarly adverse to mucking with xorg confs.
Let me know if you get any progress
Hi there! Your text contains links to other Lemmy communities, here are correct links for Lemmy users: [email protected]
Apparently Ubuntu & Fedora support multiseat out of the box. I dont know much. But I guess that could help your research.
From what I’ve heard, true multiseat is very to configure. You probably also want to investigate using GPU accelerated containers, because it’s legitimately easier to share the same GPU across multiple containers as opposed to multiple seats.
