I’m looking at setting this up soon, and read in the Github readme that if the master doesn’t have any of the device type added that you’re trying to pull over form the slave, then you’ll need to add the blank domain to your config. So if you’re slave instance has switches, but the master doesn’t, you’ll still need to add switch: without any switches to your configuration.yaml. hope that helps.
Another option I’ve seen, and am debating about using is the usbip linux utility. It passes usb devices from host to client via IP. I’m not sure which would be more reliable.
Thanks I noted that but it wasn’t the issue. In retrospect it may have been that I didn’t do something with ZHA on the VM side.
I’m now using USBIP with my Home Assistant Supervised on a generic Linux host (HASS in Hyper-V) setup. It works! However:
The service was not initializing correctly at reboot on the home assistant side. My first thought was that the issues was it was running the service run before network connections have been established. network-online.target didn’t work but
Hi my Home Assistant is installed on my ESXI using the VMDK virtual appliance, how do I install “socat” if I make a putty I only have the Home Assistant terminal.
But, when you restart HA from the interface itself only the docker container for homeassistant restarts and not the supervisor service, right? So, socat will not restart when you use this approach and restart from the webinterface, only if you restart supervisor as well?
Assuming you use socat the way I use usbip, yes, correct.
Fortunately I don’t lose wifi connection to my PI ever so I don’t need to do this.
Unfortunately the windows 10 hyper-v host I have Homeassistant on force restarts itself every so often.
So for the last few months I’ve been successfully running home-assistant on x86 machine that was remotely connecting to a z-wave usb stick on a RPi, orchestrated on Kubernetes. (however, if you just want the socat commands, you can just extract it from the kubernetes manifests below)
The new ozw-zwave-mqtt support that’s being added will obsolete this (and I’m so excited about that!), but this recipe may still have use for folks who may want to do this for zigbee or any type of other serial interface on a remote machine.
My configuration:
Kubernetes cluster (I’m using k3s+metallb and nfs server for persistent storage).
VMs on ESXi running as my normal x86 workers (home-assistant will get scheduled on this)
zwave-server (this will expose the Z-Stick as a tcp service)
zwave-client(this will connect to zwave-server and expose it as a pts device on the k8s node)
home-assistant(this will mount the /dev/pts so it can access to the client device)
I’m using nodeSelector to place zwave-server on the RPi, and then colocate zwave-client and home-assistant on the same node. (The reason the zwave-client and home-assistant are two deployments is so that I can run the zwave-client in priviledged mode, and keep home-assistant locked down)
So my situation is the following:
I have Home Assistant running on a Rasberry pi 4B 4GB.
And i have 2 rasberry pi zero. And i want to connect a zwave and a zigbee usb to each.
And then place the 2 pi zero inside the house. And the home assistant is in the garage 30 meter away. Whats the best way to connect these pi zero’s to my main home asisstant. I’ve read this guide but it says it needs to install some apt on the rasberry, but i am running the native iso from home assistant. Can anyone give me some insight?
Nothing on my HA instance looks like anything here. It doesn’t even have the apt package (or any way to install it from what I can see).
I have been doing linux admin for 30 years and I am literally lost. This home assistant thing looked like a promising replacement for smartthings but it is turning into a nightmare.
All I want is to run Home Assistant on my Hyper-V server and have it run my Zigbee & Zwave network. Is that too much to ask of this product?
Then you installed Home Assistant OS, which doesn’t have this functionality. You can run Home Assistant Container, Home Assistant Supervised on top of Debian or Home Assistant Core in a venv. Here’s a good overview of the different install methods and suggested skill levels.
That’s definitely possible, lots of people do this (including myself). It’s better to open a separate topic with your issues.
I don’t know what you are trying to achieve, please open a separate topic for your issue and give some more details about your goal and current setup. You can tag me there if you like to.
I keep getting referred to this post for usbip, but I’m using Debian 10 and it won’t let me install linux-tools-generic – has anyone successfully gotten usbip installed on a debian 10 instance?
Yeah, it looks like it’s not a part of the current Debian repository. Looks like there might be a way to install that package by adding a different repo : here, but I’ve just been manually mounting on restarts until I get it sorted:
this has been working well for meI even have a udev rule reattaching if it gets disconnected, until i added another instance on both client and server for a bluetooth dongle. I noticed that the script as i understand it won’t work with a seconds instance as i understand it on the client the service stop will pick the first port usbip port and detach it regardless which port is actually required to be stopped. IYSWIM…I have also a problem with starting it also where it starts the service ok but just doesn’t work properly…this requires more investigation (starting manually works ok)…
Hi there. Anyone finished a working udev rule for usbip host? Have trouble to get one working on RPi. Second thing is a reconnect script on the client, any ideas?