I am running OH since years now with a ZWave Binding. I recently updated to the latest version on openhabian so that the new Shelly Qubino Roller SHutter item will work.
After including the roller shutter item, I started to realizes issues with the network.
I got a lot of messages that zwave messages are rejected by the controller (see the attached log file).
The only item I can control is node 2, which is also a roller shutter. All the rest is not reachable anymore.
If I shut down openhab.service and start the zwave-js-ui, I can see all nodes in there and also control all roller shutters. But also there I am facing an issue that when a node wakes up (I got several window / door contacts included) ZwaveJS wants to start the identification process and this one is failing then also (message queue full).
I am totally lost what the issue may be here. Anyone got any idea? zwave_log.txt (980.0 KB)
I wish I had better news to your issue. The only thing I can add is that everything seems normal until a message Powerlevel_Get with node15. What does the ZWJS network look like? Could be a controller issue. Is it near a source of interference? Do you have a spare controller to copy a backup to?
There is a difference in how the OH ZW binding and ZWaveJS startup. OH ZW requests a NIF and then asks about dynamic values, so there is a lot of traffic at startup. ZWaveJS just pings the powered nodes and waits until new messages arrive (some values could be stale for a while). Anyway it is not surprising why the problems appear later with ZWaveJS.
thanks for your answer. I was now able to narrow the whole thing a bit more down in zwaveJS. In the moment where I request a re-interview of any node I receive an error and then the whole network is broken until I soft-reset the stick. Your explanation that openHAB does this interview right at the beginning for all nodes explains, why the network is failing there so fast…
So I assume it could be a hardware failure of the stick?!
I got the zwave.me stick
Bus 001 Device 077: ID 0658:0200 Sigma Designs, Inc. Aeotec Z-Stick Gen5 (ZW090) - UZB
Can I copy my network over to another stick (which I need to buy first), or is this too risky because of this failure?
In case yes, what tool should I use for such a stick? Can I use the NVM-Management (Backup / Restore) from zwaveJS? Do I need to have exactly the same zwave stick, or could I use any?
However, the controller rejecting is generally because it is very busy. I’d look through the statistics in Zwavejs (assume you are using the UI) by node to see if there is a bad actor or two. it may or may not be the zstick.
As to the transition, the nvm-managment is wavejs is an option. I still believe there is a compatibility issue when moving from 500 to 700/800 controllers, but I would ask as a discussion on the zwave-js github. I know they were working on it.
so, I now received a new stick and managed to do a backup of the old stick and import it to the new stick.
Now, with the new stick attached, everything is running again. So I assume that the old stick was corrupt somehow. Hoping that the issue will now not raise up again.
