Hey guys,
I have hass.io installed on a docker inside an ubuntu VM, inside Server 2016 Hyper V… But i’m wondering if this is not overkill?
Easie enough just to have a dedicated VM for home assistant.
So whats the best set up for this usecase?
IMO, docker is the easiest way to run and maintain home assistant
I’m currently running Hassio inside docker on a Debian 9.5 VM inside a WS2k8R2 Hyper-V host.
I believe a Hassio VMDK is also available to directly use as a virtual appliance.
I’ve seen some posts on this forum about folks upgrading from RPis to NUCs to improve performance, so perhaps it is better to start from a system that is overkill instead of starting small and having to upgrade later. (although I am very interested in buying a RPi just to play around with it)
Thanks for that info.
Apparently it does not run in Hyper V.
Anyway, I have it working for now 
I have this exact setup as well, if you think it’s overkill you can always install Blue Iris for the camera part and watch your CPU usage skyrocket 
It seems to be the way it’s done lol.
I have Arlo Cams, but they dont seem to work 
I did the pretty much the same,
Main HA installed on NUC,
Secondary on HyperV for face recognition (updates main HA via MQTT) that runs on Windows 10 PC with BlueIris.
I have 20 cameras recorded in that PC @1080 - CPU gets to 40% top.
I’m thinking about going from virtualbox to hyperv. Does Hyper V let you pass bluetooth devices through from the host to the guest OS? If that’s the case, is the device being passed through still available to be used in windows?
@vermis No you can’t. Instances running in Hyper-V are isolated and don’t know about a guest OS or a guest hardware.
1 Like
No native solution, You can use 3rd party software - usb over ethernet or usb redirector to do that
1 Like
Heh, it figures I also have 16GB of RAM, but 2 cores and no HT definitely don’t help…
Same here Ubuntu VM in Proxmox, Homeassistant and everything else in Docker containers. Best part: Traefik as a reverse proxy. Running on a NUC i5.