I am and I will continue to. I switched before they crippled Z-WaveJS so I really did not know they eliminated the ability to use tcp:// until other users could not input it like I had when it first came out…
I think someone else here set up USBIP. I think that creates a local tty port?
Is that broken now as well?
Did you reset then add those devices from scratch, or were they added before the update ?
Added before. Scroll down past that post. Someone else asked how to get them to register properly.
“running the Z-Wave JS server remotely is a way better choice and experience, which we should promote.”
This is exactly what I set out to do on my RaZberry which needs to run independently of HASS. But when I dug into what I needed to do, I saw this:
Run the Z-Wave server yourself
This is considered a very advanced use case. In this case you run the Z-Wave JS Server or zwavejs2mqtt NodeJS application directly. Installation and maintaining this is out of scope for this document.and this:
Z-Wave JS Server
These instructions are for development only. These CLIs will be available as zwave-server and zwave-client after installing the NPM package.That scared me off from running the server remotely, and found it was straightforward to run ser2net and connect via tcp. I just had to uninstall Z-Wave-JS and go with the MQTT variant to accomplish it.
My thought was to keep the existing device pick list as-is along with device=tty validation, but simply add a new option to enter a network device with or without validation and problem solved. If in fact installing the server independently was actually promoted through documentation that showed how that wasn’t “for developers only”, well, that would have been nice.
It was just pointed out to me, why not fire up a barebones HA on the Pi just for the HA Z-WaveJS addon? Then the integration on your main HA would connect to ws://[ip-of-pi]:3000 rather than ws://localhost:3000?
Possible??
Of course, you would lose the nice MQTT control panel until such time as HA incorporates it, unless they decide to deem that an advanced use case…
Why not just run zwavejs2mqtt in a container? Or run just the server directly on the os? It’s a few lines of cli and it’ll do the same thing
Anyone updating entities name for better and worried about historical data in influxdb?
Ease of maintenance/upgrades? Log into bare-bones HA, Supervisor, update, done.
1 Like
Yes, was forced to since they all changed anyway!
Nope!
1 Like
@dbrunt - I have other services I need to run to connect to a security panel which aren’t supported by Home Assistant through any integration or add-on. And ideally, I would like to keep using the RaZberry/Raspian image with Z-Way server that I can toggle off and on as needed for firmware updates or whatever.
I think the best approach is to just run the Z-Wave-JS server directly on the RaZberry Pi, but as my comments above showed, I wasn’t sure how.
@Petro - “just a few lines of cli”. Nice. Can you point me to a doc please? I found this:
https://zwave-js.github.io/zwavejs2mqtt/#/getting-started/quick-start
but there are several options. Most seem to require I install Docker first (on Raspbian, in my case). And all the examples install zwavejs2mqtt, but I need only the server, right?
I"m using the USBIP and I never used port 4000. I’m using docker for my zwavejs2mqtt and it exposes the websocket port my HA talks to it just fine and my server is none the wiser that the USB is actually mapped over the network.
Any document link please about how to guide for USBIP?
Basically looking for docker option if possible to run USBIP.
I have already container for zwavejs2mqtt and HA(Core) but have not switched to it and still using native zwave.
I run everything under docker on Ubuntu.
So this is how this will work. Say you have a pi somewhere in your house that is more central located than your HA server. You’ll plug your dongle into that following this guide
https://derushadigital.com/other%20projects/2019/02/19/RPi-USBIP-ZWave.html
Its going to map that USB device back to your server. As far as your docker setup goes nothing will change because you essentially just created a virtual USB port on your HA host. You’ll still specify /dev/ttyUSB01 or whatever and it will work like magic.
1 Like
What the advantages of this method over running the zwave-js server on the RPi instead?
An advantage of running the server on the RPi is that it doesn’t shutdown the network if your HA host has to restart. With this remote serial port your network goes down.
1 Like
My case was special. I have a nortek dual stick and the zigbee2mqtt does not support the nortek so I needed to use the local ZHA integration for my handful of zigbee devices. I initially tried to do it all on the pi :). So this still got me what I wanted but in a round about method.
So ZHA is like the “classic” zwave integration. That’s an interesting use case, thanks.
no sorry ZHA is the classic zigbee integration for HA its built in like how zwave is/was before OZW and this new hottness
Yes, that’s exactly what I was saying. 
1 Like
Does USBIP establish/create/connect server1:ttyUSB port over to server 2:pseudo-ttyUSB port and you then specify the local-psuedo-ttyUSB port in Z-WaveJS-MQTT? What I’m wondering is whether the new Z-WaveJS addon will recognize the pseudo-ttyUSB port…
You should be able to install HA on the Pi possibly by utilizing the…um…gee, I forget the procedure now. I’m thinking it would be like the one I used to install HA Supervised on my NUC under Debian before I got this HA Blue with it already loaded. It would just be an add-on to what’s there now so you’d still have Raspian OS, z-way-server, etc. Installing docker on the Pi is straightforward.
I sounds though like USBIP might be the simpler approach, if HA Z-WaveJS addon will use it. I know OZW would not…