I did something similar. Everybody is different so maybe not the best option for you, but who knows.

I have a single mini PC that handles my stack of virtual machines hosting various things. For the main OS I went with Fedora KDE. I chose something with a GUI for two reasons, the primary being that sometimes… Maybe not as often as you get more familiar, but sometimes there is an easier way to accomplish something in the GUI than in the CLI. Things like system settings. You can save a lot of time looking up commands and syntax by flipping a switch in the settings application.

Second and most important reason for the GUI, I watch TV on this thing. Which I would not recommend if you are hosting anything that can’t handle a little downtime. Once in a while a web browser may hang, bluetooth could fail, and you end up having to restart. Nothing I host is critical to anybody so this isn’t a big deal to me. I also find a little inner peace knowing that I am interacting with the main system controlling these hosts on a daily basis. If it does get compromised in some way this makes it just a little more likely I will notice quickly.

So that’s the hardware system and I’m running Libvirt as the hypervisor. It’s pretty bare bones, but easy to use and gets the job done. Hardest step to me was generating SSH certificates/keys. Not that it was hard moreso just new to me. Libvirt will not allow you to connect remotely with plain text. So regardless of your threat model this is a required step if you want remote access to the hypervisor remotely.

If you make it that far you can start really getting into the weeds with networking. I’m not going to go into the topology of my network, but I will say if you are hosting anything public you should do as much as possible to isolate that from your home network. You can create a VM to act as a firewall/router for other VMs.

Teamspeak is self-hostable and has streaming.