Sorry, no idea. Since it happens with multiple devices, I would guess a bug on the controller side.
This has just happened to me. Out of nowhere a group association to 255 was added to both an Aeotec Multisensor 6 and an Aeotec Microdimmer on my network. I had an existing, trouble free zwave network that had grown over 5 years to 16 nodes:
- Aeotec ZStick Gen 5
- A few Aeotec Microdimmers
- A few Aeotec Nanodimmers
- A Qubino low-voltage shade controller
- Three Aeotec Multisensor 6
- Two Aeotec remotes
I had just made a number of changes:
- Disconnected an Aeotec Nanodimmer
- Installed a GE/Jasco ZWave switch (ZW3010 / 46203) in its place using secure add from HA
- It was an unknown device, but I found that it had been added 5 months ago to open zwave
- Stopped HA, ran a pip install --upgrade of python-openzwave, but that didn’t have the new XML
- Put the new XML into the proper directory and added a link to it from the manufacturer_specific.xml
- Deleted the switch’s Node definition from the ozw config XML
- Started home assistant
- Ran a network heal, since the old Nanodimmer and new GE switch were in a somewhat central location and the zwave network seemed laggy and somewhat unstable
At this point I thought all would be well. But the laggy/unreliable network problem did not go away and I was seeing some callback errors and transmit timeouts in my OZW logs. I think the “all lights on” problem had begun but I hadn’t identified it as such since it seemed like it might just be delayed commands finally getting through. Upgrading to latest home assistant seemed like a good idea, in case my upgrade of python-openzwave had an incompatibility with the older one I was running. Unfortunately I ran into the UI becoming unresponsive and generating error messages that others seem to have reported with 109. I then rolled back to 108.9. I ran another network heal. I also observed that the Nanodimmer that was sitting disconnected on my table was not marked as dead in the zwave network.
And then the pattern of “all lights on” followed by “all lights off” problem became quite apparent as it repeated over time. There was no log of the lights on/off commands in the Logbook, the HA log, or what I could parse out of the OZW log. After this “all lights on” problem continued into the night I started googling and found this thread. Looked and found the two devices with 255 group associations and cleared them. Problem solved for now but concerned about how someone else said the associations kept coming back.
I’m really writing to confirm that this issue has been seen again more recently and provide all of my activities leading up to it in case it causes an “ah ha!” moment for anyone more knowledgeable about zwave and/or home assistant.
My next move is to reconnect that nanodimmer elsewhere in my network and re-run the heal. I did not want to remove it because I knew I would use it again soon and generally avoid remove/re-add cycles since I’ve read others saying the node number always increments and when you hit 255 you have to reset your whole network and re-add it all again, which sounds terrible and I hope never to do.
This is false, it will cycle thru any unused node numbers after 254 (i recall 255 being reserved).
As for your main issue, the way zwave works is if the network controller goes down for whatever reason the mesh network keeps going.