Why are thermostats always missing after startup of ZWave JS?

I have had OZW running for 3-4 years, without any big problems (a few times I have had smaller issues, but they’ve always been resolved). But after going to ZWave JS to MQTT I have this problem where a restart of ZWave JS invariably will cause devices not working in the ZWawe JS to MQTT control panel, and they will be missing in Hass too. It seems to be the Heat it! thermostats that are mostly affected, but I’m pretty sure I have seen it on switches too.

The missing devices will be marked “Status Unknown”, and “Interview complete”. Even after several days they won’t show up. If I run a re-interview from Hass or a ping from the ZWJSMQTTT control panel it will be shown as “Status Alive” in the control panel and up and running again in Hass. What could be the issue here? None of them are moved, none of them have any connection problems.

It would be easy if I could run a ping from Hass,then I could automate it on startup. But from Hass I can only do a re-interview, and that uses a bit of bandwith with my 22 devices, I believe.

Not sure what the problem is with the devices, but there is a ping service call if you want to automate something.

Ah, thanks! I didn’t find that, but I guess I looked the wrong place. Great, then I can just run that after each restart of Hass. :+1: :metal:

No wonder I didn’t find it… It’s not there. At least not in the services of my Hass. I tried Developer’s tools, but no service like that excists:

I will update Hass tomorrow and see if it is something added after 2021.6.5 version.

I had time to upgrade, and now I have the latest version. New items are ping and refresh values. But they don’t work. I have tried to run Ping several times, and refresh all values once, for the Thermostat that was unavailable on this restart (lucky number 13 this time…), and it doesn’t change to show values. I selected climate.thermostate.13 in the UI, so it should be the correct one. There are no errors in the Hass log. I have used the UI and the YAML versions, nothing helps.

The weird thing is that now I can’t get it to work doing pings from the control panel either. I have done a reboot, and it still doesn’t work. It seems like pinging from Hass stops pings from the control panel and blocks the entity completely. After a reboot I tried with another thermostat, that was working before, and now I can’t get that either in at all. I am working over vnc, but I need to try to pull the plug of the Pi tomorrow and see if that helps or if I actually have to wake the thermostats physically. That would really be a problem for me.

Edit: Reinterviewing doesn’t do anything either! But I managed to get them back by using healing…

Edit 2: I probably should tell the developers about this at Github, I just f’in HATE those long winded “how man cups of coffee did you have before this happened” forms…