Most nodes dead after OS and Core update

Configuration Z-Wave
1

My Z-wave network has been running (mostly) without issues for several years now. I upgraded to the newest OS and Core the other day and now, all but two nodes are showing up as dead. One of the working nodes is the controller, the other is the newest switch to the network, which was added about 6 months ago. This switch is also the only node on the network using S2 security.

I’ve tried healing the network, multiple reboots, even moved the system closer to one of the dead switches to see if it was a transmission issue. Nothing has worked. Any ideas?

Hardware: RPI3, Tricklestar 300ZW Controller
HA OS: 9.3
HA Core: 2022.11.4
HA Supervisor: 2022.10.2
ZWave JS: 0.1.74

2

pull the controller out of usb for some seconds and stick it back. That did resolve my issues already twice in the past.

1 Like
3

Didn’t fix it. Pulled it out, waited 2 minutes, plugged it back into a different port.

4

i don’t think, that helps.

5

Didn’t make it worse either. :slight_smile:

1 Like
6

Turns out the issue had nothing to do with the OS and Core update, just coincidental timing. My thermostat was having issues that somehow messed up most of the Z-wave network.

Based on some other threads, I decided to go to each node and test it for malfunction. First stop was my thermostat. The low battery indicator was on. I changed the batteries but the indicator did not go away. After messing with it for a while, with no luck, I did a full reset. Reset not only fixed the battery indicator, but now the whole Z-wave network is working again. Who knew?

1 Like
7

You may want to work on strengthening your mesh. If all comms to devices go through that node, then a failure of that node can take down the whole network. Do you think that is what happened?

8

I don’t think this node is a choke point. While it is the closet node to the controller, the other node that continued to work is actually closer to most of the other nodes that were not working. While I did not check the logs, I assume that the working node would have been better able to communicate with the other nodes.