I’m currently migrating my z-wave devices to Home Assistant (using the current outdated native integration).
For the current test I’ve purchased a Z-wave stick and migrated 1 powered switch and 2 battery powered sensors.
As I added the powered switch last I know the controller can reach directly the two battery powered devices.
I’ve tried to show my network using the z-wave graphing solution, but it shows all my devices as unconnected. I don’t know why.
If I look at the ‘neighbours’ attribute of each device and manually draw a map the situation is supposedly as follows:
So the controller is only connected to one of the battery powered sensors.
This cannot be true as I can perfectly control the powered switch.
I’ve tried healing for example the powered switch to get to mention the controller as a neighbour. In the logs I see the error “WARNING: REQUEST_NEIGHBOR_UPDATE_FAILED” but I also see a message
“Assigning return route from node 4 to node 1” where node 4 is the powered switch and node 1 is the controller.
I notice for the battery powered sensors that occasionally the Home Assistant doesn’t get the new status of the sensor, (if I open/close a door, it doesn’t always get reported) so I’m trying to understand the problem…
Bedankt voor je antwoord!
I assume you refer to the distance between the controller and the powered switch right? The distance is about 4 meters.
All devices have the “Is zwave plus” attribute set to true, except for the controller, where it says false. This is very strange as the stick I just purchased is the recently released Aeotec Z-stick Gen5+, which most definitely is a Z-wave Plus device. Could that be my problem? I have the impression from this post that the device is supported. I know @gcoupe is using that new stick as well, but I believe he is using the qt-openzwave integration.
Correct. However, the Aeotec controllers (both the Gen5 and the Gen5+) also show ZWave Plus capability marked as False in qt-openzwave. I think this may be a quirk of the software not correctly reporting the capability of the controller.
Based on @RickKramer 's question about the location of the powered switch I did a little test and placed tat switch right next to the controller. I also did a network heal on all of the nodes and woke the battery powered ones up. The result is now in the graph that the devices are connected:
I understood that with z-wave plus it was supposed to automatically find the neighbours and no network healing etc was necessary?
Also clearly it was an impossible situation that the controller was able to control the powered switch without it being a direct neighbour (as it’s the only powered device). But then again I seem to understand that in practice it doesn’t really matter much whether a device is listed as a neighbour or not for the actual communication between the devices?
The controller doesn’t report if it’s zwave plus or not, it’s a property of the nodes. The real bug was probably reporting this for controllers in the first place.
The dev branch of ozw addresses this by checking for other properties that infer the controller supports it, basically faking it.
