Did you restart HA after you selected the node and used Remove Failed Node?
Hello,
thank you for your answer.
Yes I restarted HA but unfortunately the device stays.
Is there another way to delete the device ?
Not really. You can factory reset the Z-Wave stick, but that wipes all the devices from the controller. The other option is to ignore it by using the control panel to exclude the individual entities.
Thanks for your advices! 
I just digged into this again, and I found this:
I have two Telldus - plugs, node 12 and 13… The new_entity_id and old_entity_id are equal. How do I adress those nodes in a configuration without pointing to the node id?
Thanks!
//Sam
Do you have
zwave:
[...]
new_entity_ids: true
in your configuration.yaml?
No I havnt got that… Shuldn’t t be opt-out?
I’ll try!
The screenshot shows your actual setup? If so, you can see that the old entity id has an added _12 or _13 representing the node id - and the new entity id doesn’t.
Since the names are identical otherwise, HA starts enumerating them and you get telldus_..._switch as the name for the first device and telldus_..._switch_2 for the second one.
To make your life easier, you should assign more descriptive names to those nodes using the “rename node” function from the zwave configuration panel. For example, you could call one telldus_kitchen and the other one telldus_livingroom.
Since the names then differ, HA won’t add any numbers to it and you will always be able to refer to the device using this name.
Sebastian
Thank you for your answer.
It’s a workaround and it could be fine that the functionality remove failed node works :slight smile:
Ignore them. They’re attributes that don’t matter at all.
I had 4 nodes that were bogus, they were from previous failed attempts to include devices. I had tried to remove them in the past with no success. I decided to try again today. I did a “remove failed node” on all 4 of the bad nodes, then restarted home assistant through Configuration–>General–>Restart.
Only 1 was removed.
I repeated the process with the remaining 3 nodes, “remove failed node” on all 3 followed by the same restart. Another node was removed.
I then repeated with the remaining 2 nodes, again one was removed.
Now I have 1 bad node and it won’t seem to go away.
1 Like
Here is what I tried and works for me. But it’s a pain in the a**.
Go to Configuration -> Customization -> your device (zwave)
Change the attribute is_failed to true.
I also change the attribute is_ready to false and is_awake to false
Restart HA.
Double check if HA takes these changes.
Go to Configuration -> Z-Wave -> Entity : Your device
Hit Refresh Node, if it take more than a couple of seconds to execute, then it’s working. If it finished immediately, not good. double check the attribute change.
then Hit Remove Failed Node, it should take a moment.
Hit Refresh Node and Remove Failed Node again and again until it finishes immediately.
Restart HA
The bad node should be removed from HA.
Not sure why it’s such a pain in the butt, but I’ve tried it 3 time in my pi and it works.
11 Likes
I managed to remove the node also by just trying and trying again to remove the node. It’s finally gone, but in the Z-Wave integrations there’s still an entry that I can’t get rid of. I’m guessing it’s a db entry and that there’s no code to clean it up after removed nodes.
I have this issue too @seanodell …is there any solution to it ? Have you tried deleting the integration and allowing discovery to find it again ?
More or less. It seems like HA gets its list of Z-Wave devices from the connected hub, but that there is a bit of caching somewhere, though I haven’t spent any time yet to figure out where.
Here’s what I do now to get rid of a device:
- Perform a “Remove Failed Node” through HA itself.
- Completely power down HA and the Z-Wave hub itself. I’m running Hass.io on an RPi so I turn it off and on again.
I managed to get that working with a “dead node” on the first try, but previously I was removing the failed node and restarting HA itself. Apparently, just restarting HA is not enough. Everything needs to go cold first.
That said, the entity_id that I gave the device is “remembered” by HA so, even though it no longer shows the Z-Wave device, if I try to give any device that same name, it will complain that it’s still in use.
1 Like
Delete the old entries from core.entity_registry in the storage folder and you’ll be able to rename your device with the old name. Make sure you get all related entries and don’t mess up the json format or you’ll be screwed.
Do this at your own risk.
This worked perfectly! Thanks.
I was having issues removing a couple z-wave nodes. I tried removing them both at the same time and it wouldn’t work. However when I tried removing them one at a time, restarting home assistant after each one, it worked.
Thank you. This worked as described!!!
This method worked so it is still only working solution I found in 2020. Thanks.
I have a zwave node that corresponds to a device that is no longer functioning. This means that I cannot for it to be removed with the standard procedure (setting the network into Remove Node mode and physically interacting with the device itself).
I tried to simply remove it from the core.entity_registry and core.device_registry, but as soon as I restarted my HASSIO, the device and its entities were there again.
How can I permanently remove it so that it is completely forgotten the same way it would be if it was removed as I mentioned above?