so finally with pci passthrough enabled i was ready to boot the vm. i had already downloaded official arch installer iso, set that as the iso and booted the vm. arch is an intimidating install, i worked through it a few times. best way is do a snapshot backup right after it boots so you can go back without having to destroy and recreate the vm. eventually you get the root arch prompt. now there are a ton of guides out there, but for me the easiest was to ping google, make sure i had internet and type archinstall > enter
this is by far my preferred install method and takes alot of the headache away…towards the end, select the graphical wm you prefer – i use gnome and so i went with that. i tried alot of different methods, but if you just want pci passthrough working the graphical installer will pull the correct drivers. if all goes well, everything will install and let you get to a point where you need to reboot. do that, then manually stop it, disable the iso and start the vm again and you should have a shiny new vm with working graphics drivers. at this point, do a quick lspci and check that pci device was detected and if you are in the gnome desktop graphics is indeed working. this is another great place for a backup (snapshot), install a few other packages, these vary in the names between distros but i used nvidia-utils for arch, you can also install cuda. once those are installed a nvidia-smi command should show you that things are operating correctly and then you can install jellyfin-web jellyfin-server and jellyfin-ffmpeg. start and enable the services using systemctl enable jellyfin and systemctl start jellyfin as per standard practices so they boot, then navigate to the default <your IP>:8096 to initiate the server… next was getting arch and jelly to see my files via nfs…