Odd issue with 70.0, entitty_id registry, and zwave

I found some curious behavior that would be nice to get some insight on.

The situation:

  1. I have a Zwave outdoor GE outlet. It’s one of those inline plug ones. I use it for Christmas and halloween lights.

  2. In entity registry it’s set to have a name “zwave.outdoor_1” with a corresponding switch “switch.outdoor_1”

  3. These two relate back to the zwave node ID and work correctly.

  4. When not in use this outlets get unplugged. During that time it’s reported as a dead node.

  5. But what is odd is the switch corresponding to it; “switch.outdoor_1” seems to randomly change state over time. This is odd as the physical device isn’t available.

  6. If I plug it back in after a few minutes it works as expected and the status for the switch matches the actual state.

  7. If I remove the switch after several hours I start to get the phantom state updates.

To me this seems like some sort of odd un-intialized variable issue; or memory being overwritten to cause the state change. I even see the state change in the logbook when it happens.

Does the state change show up in the OZW_Log.txt as well?

Be careful, at some point it may get marked Dead within the controller, at which point it will refuse to communicate with it.

When plugged in yes, but not when removed. It is obviously a ghost state change,

Yes, worried about that - any suggestions for how to do this without leaving them plugged in somewhere?

Leave it plugged in near the same location, so that the mesh stays healthy, or move it somewhere else and run Heal Network.

Or, remove it from the mesh, and then re-add it the next time you want to use it.

For a well-behaving node being “Dead” is not a problem. When you plug it back in it will send a packet to the controller and the controller will consider it “live” again.

Do you have a poll configured for that switch? Either a z-wave poll (which should be fine) or some other automation?