How to remove ghost/unknown/dead Z-wave nodes

Please don’t see only the route numbers, it says “route unknown” right in front of it, because Homey lost connection to the device (also why it says “unreachable” under flags), it just doesn’t know the route.
Try giving these devices a “test” command first (under the 3 dots behind it), that should resolve it, if the device is in range, as it is definitely possible that it is just out of range.

Sorry, I cannot follow your reply completely. I think you are stating that Homey is not to blame and has no influence.
However, there are options to resolve these problems on other platforms/tools. Take a look at https://community.openhab.org/t/remove-a-ghost-z-wave-node-from-habmin/6640/6

Being out of range might be the case (a bit), but there is absolutely no reason whatsoever to include node 16 in any route… since it doesn’t exist.
It might be that it has been a malfunctioning include/exclude, but there is no reason that we shouldn’t have a tool to correct this.

The only reason for setting a seemingly random number (unused node) is either as place holder or it is a debug number so Athom can help if you contact them (only Athom knows why they needed to implement it this way), I’ve yet to see a common sense for the id’s, so most likely it is just a place holder so it doesn’t break the overview.

Trying to compare Homey with other zwave controllers is like comparing a Mercedes with an Audi, they are both cars, that kinda work the same, but Defenitly don’t work exactly the same.

Homey sets devices as “unreachable” when it didn’t receive an acknowledgement back from the device for 3 times (that it is alive), making the last route unkown after a reboot/certain time of no communication.
You can initiate another alive check with the “send test” option, making the device possibly reachable again.

Please allow a question that doesn’t quite fit the topic.

Can you please explain why I a have a lot of main power operated devices there are mark as unreachable again and again. But they worked all time without any issues. And after press on Test they are reachable again. I thought that homey doesn’t poll
the device periodically. So it can not happen then a main power device will be marked as unreachable?

Homey by default indeed doesn’t poll devices (though that can be implemented by the developer with just adding 1 short line of code).

I have noticed myself that certain devices (only 2 in my case) that gets put into “unreachable” after a certain amount of (random) time, not sure what the correlation is as to why it is put into “unreachable” even though it works just fine, and also still shows the “last working route”.
And because it works just fine, i don’t really bother to investigate, and ignore it.

As mentioned lots of times before, your node is flagged “Unknown”. This means that Homey can’t communicate with it for whatever reason (empty battery, power disconnection, range or interference issue). Homey therefore also cannot determine the last or the current route and it does sometimes show a strange route. That’s what it is, nothing more nothing less.
You should determine why your device is set to “Unknown” and solve that issue.

@PetervdK: To my knowledge Homey has no influence on the routes of ZWave devices. Routing is completely embedded in the system/protocol. Homey doesn’t ‘randomly’ add nodes, my node 16 is somehow still in the routing table of my ZWave network… this is not Homey’s fault, however might be something Homey can fix.
The unreachable node (27) is right next to a node (36) that is working fine so ZWave should reroute, or Homey should give a command to reroute.
As pointed out in the other (openhab) thread that I shared… there are ways to fix this, only Homey cannot (yet) fix it.

They only thing I want to know is how to remove (ghost) node 16 from my network, that is all… no more, no less :slight_smile:

Healing and testing does not work. When I do a power-cycle of the affected devices everything works again. But for how long :slight_smile:

Read my posts (not sure if in this topic or elsewhere) about placing devices too close to each other. That tends to cause problems with the mesh, causing delays and unreachable devices.

Well, good luck with dealing with the effects instead of taking care of the causes :+1:
Unknown is unknown, whether you get rid of the ghost node 16 or not. In fact, if all your nodes were occupied and working, there would have been another fake and non existing node number there.

I have tried to remove a fibaro Zwave powerconnector and now it became “unknown node” if I test it in Zwave it works fine but I wanted it to be removed.

Now I cannot use it anymore as this device is not visable in flows anymore so I cannot choose it anymore. Any help with this ?

Did you read the topic? The exact solution is in this topic a bit back…

1 Like

Already fixed it, it was really very simple, just unplugged the powerplug from the wall send a test zwave command it did not reach the device anymore and then the greyed out REMOVE became white.

Would have been nice if this is documented somewhere is this is not a straight forward thing.

@PetervdK I did remove the device through the gui but it was still not removed.

So now I put it without power and then send a Heal or Test command which ungreys the remove option.

So all set now and have readded the device to homey again.

Has this ever been sorted out? I had some routes through non (even never) existing nodes as well.
Yesterday, I have removed my Greenwave power strips from my network. I still use them, but as dumb power strips now. However, the ID of one of them is still noted in the “last working route” of 6 Z-Wave devices. The devices do work and are still communicating with my Homey.

I have done this in an attempt to improve my Z-Wave network and force end point devices to route through my Coolcam power plugs rather than using the Greenwave power nodes. Now after removing the Greenwave power nodes I still see 6 of them go through a non existing Greenwave power node, only very few take a route through a Coolcam power plug and the rest decided to connect directly to the Homey.

Have you tried a test or heal on these devices?
Most of the time this updates the last know route.

Did you remove it by removing the device from the network (with successfully removed result)?
If not, try what @Dijker says.
As an extra step you should also reset the device itself to factory settings by hand to be sure it doesn’t try to communicate to Homey anymore.

Jup, tried to test and heal multiple times. Then some of them become “Unreachable” and I get an error message. Note: they’re still working and sending data to Homey for the bigger part.

I removed 2 of the 3 power strips by using the option from Homey to delete, then push the button on the device. The other one was not responding and I removed that in a different way. However, that is not the one still being visible as part the routes. But that one I did reset to factory defaults.

Then to be absolutely sure you should do that with the other ones too.
And after that maybe PtP your Homey to “reset” the mesh.

Ok, also factory reset the one that was still being mentioned as part of the route.
Pulled the plug on the Homey and left it out for 10 minutes
Still node 42 (that one Greenwave thingie) is being mentioned.

Now, of course this is the “last working route”. So, I will wake up the devices using this node 42 a little bit closer to the Homey and see if that helps.

Ok, got rid of those “node 42” mentions in the route.

Now I have a different issue. A Coolcam power plug that is working and connecting to Homey as long as it is close to Homey, but as soon as I connect it to the washing machine (where I want to use it) it is uncontrollable and not connecting to Homey.
However, I have some more of these power plugs…the whole idea of a mesh network is that this should work through other routing devices. I will not hijack this topic for that and start a new one.

I have the same problem, and every time I have to do the same, disconnect and remove the device and reconnect it, ok for plugs and things that are easy to disconnect. But with integrated installations, dimmers, temperature controllers etc it’s a bit more work/complicated. And the biggest issue is the WAF (wife acceptance factor). As she goes mad when smart things turn stupid.
But main issue: Why does strange Z-Wave nodes appear, I guess that the HOMEY is to blame?? And why can’t there be an option to rebuild the route??
Last time I got these new nodes : 224, 234 and 244 in my system. None of them has ever been there, and I do understand that trying to se a logic in the Z-wave routing is just no point of trying to understand;)