so i tried running through this with a fresh RPI3B+ with latest no GUI distro of rasbian straight from the raspberrypi.org installer. the docker component took me two nights to get working (i’ve never used docker before). most of it was resolved by focusing on getting dependancies resolved and the specific versions needed for running docker on a pi.
i wish i kept the links for everything I did, the only ones i still have are the following.
-use this to install docker-compose:
sudo apt-get install docker-compose
then the YAML file was a nightmare to put together. i ended up phoning a friend for help on getting the path for the USB stick:
ls -l /dev/serial/by-id
returns: lrwxrwxrwx 1 root root 13 Dec 11 09:26 usb-Silicon_Labs_CP2102N_USB_to_UART_Bridge_Controller_7e8af77b0855ed119e7d93131d62bc44-if00-port0 → …/…/ttyUSB0
then plug this into the config file
/dev/serial/by-id/usb-Silicon_Labs_CP2102N_USB_to_UART_Bridge_Controller_7e8af77b0855ed119e7d93131d62bc44-if00-port0:/dev/zwave
the config file’s “example” is worthless imo. i had no idea what anything was asking for there and needed it translated to non-linux guru speak.
so yeah, a third of this post is the inquiry if anyone else had these issues, next third is how i got it working (mostly), and the last third is just a good old fashioned rant on my own ineptitude
I think most of your problems are due to following a guide that is nearly 3 years old. It just has a lot of outdated information. The project is not even named zwavejs2mqtt anymore, and the Docker instructions are incorrect.
The official Z-Wave JS installation instructions have more information, including details about Docker, minimum settings, and USB path information.
Also, unless you’re a diehard MQTT user, ignore all the MQTT instructions in that guide and use the Z-Wave integration. It’s the official way to integrate with Z-Wave JS.
Freshcoast is right. I wasted a lot of time following the guide before I just went to Settings > Devices & Services > Add Integration > Z-Wave to get Z-Wave set up. Using add-ons, MQTT, and Z-Wave JS UI proved to be unnecessary.
so here’s my problem: i’m not using the USB stick on the HA system. my HA instance is a virtual machine, so i needed something to run the USB stick on a remote device. I did come across that webpage but i’m not seeing that configuration on the ZWave links unless i’m just being dumb
Sorry, I don’t understand what you mean. Z-Wave JS UI doesn’t need to run on the same machine as HA. Usually if you are installing via Docker you are either using HA in Docker, or running ZUI remotely.
The Z-Wave integration uses a Websocket connection, which means it can talk to remote instances of Z-Wave JS, if that’s what you are referring to. It doesn’t need to be local.
my HA instance is a hyper-v virtual machine. i dont have USB access. i was hoping to setup ZWave via a spare RPI3 i had laying around
i have no idea how to setup ZWave JS UI for that scenario, nothing really says how, all the docs only reference “plug in the USB stick to the HA machine”
when i did try to do something with websockets when i managed to get JSUI running on the remote pi, i couldn’t get the instance on the HA virtual machine to start up, it always wanted a physical adapter
It will only ask for a USB path if you instruct it to install the official add-on. Otherwise it will just prompt for the remote URL. You must uncheck the “Use Supervisor” option.
yes, i’m using HAOS. and yes, i absolutely installed it incorrectly because i had nothing to go off of that made any kind of sense for someone who isn’t already familiar with the HA ecosphere
i get this isn’t a usual use case, but damn, this is harder than i thought it would be lol.
i’ll just keep fighting with it. at some point maybe i’ll write down a walkthrough for idiots like me if for nothing else documentation for my future self.
Guess I’m not sure what you’re still fighting with, after the additional instructions I’ve provided.
I’m also curious at what point you got lost in the official docs. It seems to be a common complaint that they are confusing, but it’s not obvious to me since I have experience setting it up.
Presumably, you’ve already setup ZUI and enabled the Websocket server to integrate with HA, right? That’s the hardest part.
The next step is to just install the integration. All that is involved is unchecking the Supervisor checkbox, then inputting the URL of where you’ve installed ZUI.
here’s the fun bit, i didnt find the offical docs until WAY later in the process. hence why i started, and subsequently failed at, flemmings walkthrough. nothing told me that wasn’t a good path to go down. the official docs dont even say the word remote anywhere, so in my head, there was no reason to use that doc for my setup.
i do have ZWaveJSUI up and running on my remote pi3 now, somehow i managed to get docker up and running there. i’ll try the integration on the HAOS side again and try to find that supervisor checkbox you mention. again, the doc doesnt mention anything about remote, so i had no idea i even needed to pay attention to that setting
and thank you for hearing me out on this. i’ll gladly give feedback or suggestions if i know where to send them. like i said, i have no issues making a walkthrough for idiots like me
