This is something I've been thinking about for a while; basically, I want to move my zwave node away from my main HomeAssistant system.
I'll try to be brief; my current config is a single mini/micro system (Dell, I believe), Core i5, 8G RAM and an SSD, it's a ton of power for HA and massive overkill, I know. The problem is that the system is located in a remote room of the house, so the signal isn't exactly the best and I have some nodes that are linked through 2-3 other devices; I'd like to move the USB Z-stick to a more central location, and I don't think a USB extension is going to cut it. I have ethernet wire which is far more viable to get a connection across to the HA computer. I don't want to move the HA computer away from where it is, since there's backup power where it is; so my idea would be to use something like a Raspberry Pi (now that availability seems to be improving), connected by Ethernet using PoE (for power availability from the UPS). Provided I can get a Raspberry Pi, and all the related and required parts together, which should be fairly trivial; how would I connect the zwave dongle on the Raspberry Pi to the computer running homeassistant?
I haven't considered this before due to the pi being so difficult to get since I put together the homeassistant system. Ideally, I would want several of these systems placed at key points around the house so that I wouldn't need any of the zwave nodes to relay communications, but that's future plans more than anything - I would need to source several zwave dongles and get them all on raspberry pi's and get them working together.... So going about it towards that end would be a bonus; but at least I want to do some research on it and figure out if I can even relocate the dongle at all first. Any infromation to that end is appreciated.
this is good. I gave it a once over, but I'll go more in-depth when I get the pi. Any tips on migrating my current zwave net over to it? or do I just need to move the dongle over? I want my HA to be disconnected from the zwave for as little time as possible; since the family is using it to control ~half of the lighting in the house at this point.
Right now, I should just need a raspberry pi, SD card, case, and a PoE hat, maybe a USB extension for the dongle, and I should be all set to go; so I would prep most of the software on the pi before-hand, and run the config for the dongle as soon as I take it out of the HA system. I'm hoping start to finish, it would only be ~1 hr or less of downtime.
I have had mine set up this way for 3 or 4 years now using a raspberry pi and zwave2mqtt/zigbee2mqtt and it works very well. Both of those have their own UI though which could be better but definitely get the job done. I have the pi centrally located in my house and my home assistant machine is in a rack with the rest of my gear.
I'm using zwavejs ui, which, AFAIK, is the "new name" for zwave2mqtt.
I've been using it since I started. what I don't know is how to get it communicating. It will all be on the same subnet, ethernet connected. Any tips/tricks/guides?
Hey sorry I forgot to reply yesterday. Are you talking about the add-on in hassos? How are you installing it?
I misspoke and what I use now is zwavejs2mqtt since that came out, I used regular zwave2mqtt before zwavejs and zwavejs2mqtt existed. You are right that zwave-js/zwave-js-ui are the newer tool but zwave-js-ui isn't exactly the same thing as zwavejs2mqtt even though they function relatively the same from a home assistant perspective and both are provided by the zwave-js project which is confusing and I had to tipple check they weren't the same now hah. The biggest usability difference is that to add devices to it (or the zigbee equivalent) you have to use its UI and not the integrated home assistant UI as you do when you run zwave-js and zha (at least the way I have it configured, this could be possible now). Both bridges provide discovery in home assistant via MQTT though so the experience is pretty seamless once its set up, I haven't had any issues with it.
I don't have a guide, I just run that docker container that I linked on a Raspberry Pi with the zwave controller plugged into it that is connected to the same MQTT server that my home assistant machine is connected to. I can grab the docker-compose file that I use on that machine if that would be helpful but its really pretty straight forward, I don't remember needing to read any guides that I could pass along.
Reading your post again I also wanted to mention that neither zwave nor zigbee support multiple primary controllers so you couldn't really just set up a bunch of controllers/coordinators (coordinator is the term zigbee uses). Both protocols are designed to be mesh networks and will try to mesh as much as possible, it is really key to how they work for traffic and radio frequency congestion in addition to reliability. I know it seems annoying to do but I have had great success with using a few zwave and zigbee mains powered devices (I use zwave light switches and this zigbee plug as an example) to act as routers in both meshes and it works pretty well and I think you will have much better luck if you build your mesh that way. You may also find that most controllers/coordinators aren't designed to have all devices connected to them in a network and won't perform as well when configured that way but that depends on how large your mesh networks are going to be (I have 29 zwave and 65 zigbee devices currently).
Hope that helps! Let me know if you have other questions or need more info.
ETA: I forgot to include, I use that zigbee plug even on things that don't need really need a switch but I want power metering on since they do that as well. I have one on a small mini-fridge for power metering and a zwave equivalent on my washing machine for cycle alerts a examples.