A while back I ended up with all 3 of the following installed:
ZWave JS to MQTT Supervisor Add-on (in Control Panel mode)
ZWave JS Supervisor Add-on
ZWave JS Integration
For whatever reason I’ve never been able to just uninstall #2. Everything has worked as expected, I just need to remember to stop #2 and set it to not autostart. Things have seemingly gotten out of sorts now though. #2 autostarts no matter what I do, #3 shows as fine and if I manually stop #2 Zwave stops working.
I’d like to get this setup in a more ideal situation. Many of my devices support S2 so I’ll have to go through inclusion again.
I assume my best bet is to rebuild; yuck. I don’t have any need for an MQTT broker otherwise. Are there advantages to splitting that out with ZWave JS?
In general, what is the best way to setup ZWave JS going forward?
Configuration → Integrations → Add Integration → Select Z-Wave JS → Uncheck the supervisor check box, enter the URL for zwavejs2mqtt → Click Submit. It will complain but otherwise everything will work. Now the official addon won’t be autoinstalled/run any longer.
That worked to remove the supervisor addon but zwave isn’t any better. I just turned off a light and now the whole switch is unavailable.
You mentioned the supervisor addon is official. Is that the preferred way to install it? I’m starting to think I should just get some other kind of hub at this point. I spend more time trying to fix this than I do enjoying it
Did you have the devices connected to zwavejs2mqtt add-on or to the zwavejs add-on? or both?
if you only set up the zwavejs add-on to discover your devices then they will only be available to HA if that add-on is running.
If you only set up the zwavejs2mqtt add-on to discover your devices then they will only be available to HA if that add-on is running.
You can only ever have one add-on (the zwavejst2mqtt add-on or the zwavejs add-on) running at a time since they both can’t access the same controller simultaneously.
There technically isn’t a “preferred” way to run zwavejs. It’s up to the user and their needs. you can use the “official” add-on installed by the Supervisor at set up of the zwavejs integration or you can run the “unofficial” zwavejs2mqtt add-on installed by you prior to setting up the integration. Either way will work.
There are even two other ways to run zwavejs other than those two but we’ll leave that for another time.
I did a clean switch, since I had all my devices in Fibaro HC2. But if you are using the same Z-Wave stick that should not be an issue.
What I ended up with is using only the new Z-Wave JS in Home Assistant, and since I run it as a VM in Unraid, I can just stop that VM to configure stuff in Z-WaveJS2MQTT like associations.
I installed a Z-WaveJS2MQTT docker, and use the same stick there.
I use a Aeotec Z-Stick Gen5, and it has a battery, and the Inclusion/Exclusion can be done with just the stick . So when doing the actual switch, I found that Exclusion in Fibaro is extremely difficult, and for most devices requires a factory reset. Inclusion in my house where the server is in a bunker with waterbased heating in the floors and basically not much range to include a device, was impossible (Fibaro had network wide inclusion). Going around the house with the Z-Stick in my hand Including the devices was great!
Personally, I am using the official zwavejs2mqtt from their devs along with the zwavejs integration but I am running container (Docker) not supervised.
No the devices don’t connect to the integration. They connect to the controller and are managed by the server software (some flavor of zwavejs server).
All the integration does is create a bridge between HA and the server software.
So again the question is have those devices ever been managed by the zwavejs add-on or the zwavejs2mqtt add-on or both (but never at the same time)?
I get that. Apologies if I’m coming across like a jerk, really trying not to. I’m just horribly frustrated with all of this.
The whole setup was upgraded from the now not recommended OpenZWave beta setup. When I first did that I was under the impression that the Zwavejsmqtt addon had to have a seperate server instance communicating with the usb device. I had no desire to continue to run this all as seperate containers (Hass, openzwave, mqtt) So I bought another pi and setup full blown HASS OS and plugged in the zwave usb controller. In trying to get this to work I ended up just starting fresh.
I did ended up having both the zwavejs integration and add-on installed. It is reasonable to assume that at some point in my troubleshooting each was managing those devices.
So I guess is there a path forward with troubleshooting or should I just start fresh again?
I think the best thing to do is to decide which method of zwavejs you want to run - the zwavejs official add-on or the zwavejs2mqtt add-on.
they both run in their own docker containers (just like every other add-on) so there is no difference there.
the zwavejs2mqtt has a better control panel if you need it and works fine with HA from my understanding (I don’t run a Supervised/OS install so I run mine in a manually installed docker container and I have no troubles at all). You don’t need the “official” add-on.
Once you decide what you want to use then just get rid of the other one so there is no way it will be able to try to talk to the controller.
If the devices don’t show up then at that point you can try to rescan the zwave network and rediscover devices. Again, I use zwavejs2mqtt and that is possible with that from its own control panel but i have no idea how to do it using the regular zwavejs add-on or if you even can.
Once you make your decision and get to a stable spot with only one add-on running and the other one gone and the integration installed let us know where you are and hopefully someone can help further if you are still not quite there.
You all are awesome Thanks for all the help. So at this point I am running the zwaveJS integration with the zwavejsmqtt supervisor addon in control panel mode. I just did a full restart for the host and it seems like the state of the devices is kind of all over the place.
For instance, I was just able to unlock and lock my front door (Schlage/Allegion BE469ZP) but the light switch which is currently on in the room I’m sitting in shows as off. When I sent the unlock (and then lock) command it was done from the defualt card in the Home tab (is that the right term? tab?) of HASS where things are added by default. It’s also worth noting that it took an unusually long time for both operations. I’d say 30 seconds to unlock and then another 30 to lock.
Well now it’s just gotten weirder. I’ve found a few times since I last posted that when I log into the zwavejsmqtt addon its disconnected from the zwavejs server. I also saw some logs about my locks being shared with the apple homekit bridge integration potentially causing issuses. I’m going to disable that all together since it’s not very useful to me with my zwave setup not working anyways
I’ll keep posting here as I troubleshoot but it may be sporadic as I have a generally busy weekend ahead of me.
UPDATE: (not new post worthy) I’ve commented out homekit in the config, rebooted and then uninstalled the integration. I’m now going through and re-interviewing all of the non-battery operated devices. The innovelli LZW45 light strip I’d been trying to add appeared as a very generic looking device and upon re-interviewing it is now correctly recognized. I’ll re-interview the battery devices later today and see where that lands me.
. So when doing the actual switch, I found that Exclusion in Fibaro is extremely difficult, and for most devices requires a factory reset. Inclusion in my house where the server is in a bunker with waterbased heating in the floors and basically not much range to include a device, was impossible (Fibaro had network wide inclusion). Going around the house with the Z-Stick in my hand Including the devices was great!