I summarize the actions I have taken and that seem to have worked:
I have added the virtual card as a second graphics card, keeping VNC as the first graphics card.
Previously I replaced VNC with virtual graphics, and I clarify: I was trying to connect to that VM through Windows RDP, not through VNC. Since this setting did not allow me to connect, it seemed that Windows was not starting correctly, since I could not connect to that machine through Google Remote Desktop either.
Apparently this behavior of adding the virtual graphics card without VNC, along with the GPU Statics plugin that gives me that message “GPU not available bound to VFIO or inuse in a VM” in the Dashboard, made me think that the vGPUs were not right configured. I was wrong, unless you tell me otherwise, I see everything working correctly.
My only question right now is if I should continue keeping the /dev/dri path in docker containers like the following for transcoding, or should I change it to the path of a vGPU
- Plex
- Jellyfin
- Tdarr
Thanks to everyone who has helped me!