Honeywell T6 Thermostats - all went dead at about the same time?

[Linwood-F]

This is in:

zwave-js-ui: 9.30.1.46a9eda
zwave-js: 14.3.8

I have a badly zoned heating system and control it from Node-Red by controller Z-Wave T6 thermostats.

I've had some flakiness occasionally but last night at about 9am UTC all four went "dead" at about the same time. Partial log extract:

z-ui_2025-02-04.log:2025-02-04 09:07:03.107 INFO Z-WAVE: [Node 005] Is now dead
z-ui_2025-02-04.log:2025-02-04 09:08:04.751 INFO Z-WAVE: [Node 003] Is now dead
z-ui_2025-02-04.log:2025-02-04 09:09:06.359 INFO Z-WAVE: [Node 011] Is now dead
z-ui_2025-02-04.log:2025-02-04 09:10:08.044 INFO Z-WAVE: [Node 013] Is now dead
z-ui_2025-02-04.log:2025-02-04 09:11:09.666 INFO Z-WAVE: [Node 011] Has returned from the dead
z-ui_2025-02-04.log:2025-02-04 09:12:52.227 INFO Z-WAVE: [Node 011] Is now dead
z-ui_2025-02-04.log:2025-02-04 09:13:01.197 INFO Z-WAVE: [Node 003] Has returned from the dead
z-ui_2025-02-04.log:2025-02-04 09:13:25.018 INFO Z-WAVE: [Node 013] Has returned from the dead
z-ui_2025-02-04.log:2025-02-04 09:14:06.134 INFO Z-WAVE: [Node 013] Is now dead
z-ui_2025-02-04.log:2025-02-04 09:27:41.109 INFO Z-WAVE: [Node 011] Has returned from the dead
z-ui_2025-02-04.log:2025-02-04 09:28:23.300 INFO Z-WAVE: [Node 003] Is now dead
z-ui_2025-02-04.log:2025-02-04 09:28:43.906 INFO Z-WAVE: [Node 011] Is now dead

The death and resurrection went on for quite some time. When I woke up (because the temp went from about 65 to 80 F in my bedroom), not knowing what was wrong, I restarted the docker image with zwavejs. This got most back but a couple continue to go dead and come back.

I happened to have logging at debug level enabled but the logs are massive. I am attaching a truncated portion of about 10 minutes, after the reboot, where two nodes (003 and 005) are dying and coming back.

Can I beg some help from those that understand the protocol and this software to give me a clue what is causing this?

This is a new problem: I searched logs before last night though, and have zero records where these were marked dead. The first instance began at the T09:07 time above. I was fast asleep, and there was nothing I could see that might cause it (no storms, no power failures per my UPS, etc.). The only thing I know is when the heat comes on, soon after as temperature sensors change, Node-Red is sending a lot more updates to the T6's (basically it just refreshes the proper state (heat, cool, setpoints) everytime something changes). But it's been doing that for months -- if it's a volume issue no idea why it just happened last night.

As of this moment 75 minutes has gone by since the last "dead", and 3 hours for one (Node 011) and four hours for another (Node 013).

I can't think of anything that preferentially impacts one vs the other (when I send updates, I update all four together, so they get exactly the same number of updates sent from Node-Red).

I'm hoping there is an indication of what causes this in the logs. With apologies for how thick they are, and appreciation in advance, if someone could look it would be great.

The USB stick is a ZST39LR 800 series from Zooz, and this is all running on a rPi 3B with Ubuntu 24.04.01 (current).

Linwood

zwavejs_2025-02-04.log-Truncate.txt

[hceuterpe]

In your logs, you have a nearly 40 dBm fluctuation in background RSSI on multiple channels. With spikes around -74dBm. At times, those values appear to exceed the RSSI of the devices you're having issues with.

My guess is your issue is related to interference going on?

[Linwood-F]

Hm... The stick lives on a Raspberry Pi 3B (directly in the USB).

Zwave is around 800-900mhz (right?). To the best of my knowledge I have nothing else in that range in the house. Tons of 2.4ghz for IOT stuff, some 5ghz. There's a VoIP base station fairly near the rPi but that should be about twice that frequency. A Caseta bridge is in there, though that should be half the frequency.

There is a network switch in there, and router, but no wifi, those are elsewhere (and different bandwidths). So nothing deliberately transmitting except the Caseta and Base Station.

Any thoughts where I would look for that?

I could separate the stick from the rPi I guess.

Or I could move the rPi to somewhere else in the house.

Any thoughts were I would start?

But more to this current case, any idea why it went so nuts once, when it had been running more or less in this exact configuration for months?

[AlCalzone]

directly in the USB

Not good! https://zwave-js.github.io/node-zwave-js/#/troubleshooting/first-steps?id=controller-placement

[Linwood-F]

That's fascinating, and I see it.

First, thank you for the pointer. I'll have to find a place to rearrange.

But what's more fascinating is that with several days of logs, only the day of the failure were there such large changes. Here's an itemization of today (partiall) and the day of the failure. While -99 to -112 is perhaps not ideal, it's a limited range, whereas I had -64 to -128 the day of the failure. I never saw any other times over several days below about -98.

I wonder what would cause such large change one day (in the wee hours with nothing apparent happening).

I also wonder why every other device continued to be reported and working, and only these four thermostats went insane (of about 25).

But I can (and clearly should) rearrange placement. I just need to find a place away from my network closet that I have ethernet available, it's an old house and very hard to run wires in the attic.

Thank you for looking. If you (or anyone) has any ideas why only the T6's would go insane, please chime in. Or for that matter why on one specific short period the background RSSI would go 30+ db higher (less negative) just for a few hours.

No, I'm not in New Jersey with the mystery drones.