I am using the Aeon Z-stick and has encountered a similar problem with my Z-wave entities lately. I have 30+ different nodes and despite several hassio restarts, the only ones recognized by the system are the Z-stick itself, one Fibaro motion sensor and an âuknown node 18â. I happen to have two different Z-sticks and my spare stick has been backed up with most of the nodes a while back. Plugging that stick into my RPi 3 with hass.io made no difference. Only a few of the nodes show up in HA.
I sort have have all the Z-Wave working now - 17 nodes plus the controller. For now I am just going to avoid restarts until someone can shed some light on why restarts are killing Z-Wave devices.
I can live with the renaming for now but the disappearing nodes is a killer.
Is there a simple way to just rebuild the whole z-wave mesh? All the info is alreday there in the ozw-file.
I wish there was a button to do just this⌠However my experience is that Z-Wave discovery in HA only ever seems to happen during association. With all my failures the OZWLog.txt shows devices happily conversing with the controller but HA refuses to see them.
I rebuilt my network a couple days ago and found that it dropped a lot of battery devices on reboot. This was after adding quite a few nodes in a row, so when I redid it I added one at a time making sure all the entities of the node appeared and rebooted between each device being added and it seemed to do the trick. Nothing droppedâŚ
Yesterday, I stopped HA and deleted all zwave entries from the core.* files in the .history folder. I also deleted the zwcfg*.xml file and then started HA again. After a few restarts, I could see the network slowly building up again, but not all of my zwave entities came back. So, I copied back a previous zwcfg*.xml file and restarted again. After about 3-4 hours, my mesh seemed to have stabilized, showing the following graph:
For some reason, some of my entities still donât show any connections, although neighbors are defined when looking at the entity info.
Any ideas why?
This happened to me as well. Just gave it more time, and they eventually all turned green. Give it a day or two.
I am actually considering to scrap my entire zwave network, reset my stick and start over. It seems impossible to remove nodes that are not reliable. My zwave mesh has turned into a zwave âmessâ. It would be nice if there was better functionality in HA to handle zwave entities. I am afraid, though, that if I reset my z-stick and start over, many zwave will not be willing to be included again unless they have first been excluded from its âformer homeâ, which turns out to be extremely difficult to accomplish.
1 Like
Again, I did the same 
Followed these instructions: Serious z-wave problems
I did try to exclude as many devices as I could before resetting the stick. But for all devices, I could also simply reset them and then include them again no problem. It fixed my instability issues, my z-wave network is in good shape again.
Only missing thing in the linked steps is you also have to clean out the entity registry file.
Thanks, I will probably do that soon! I have some entities (Qubino, Fibaro, Philiotech) that are built in behind wall switches etc., that are a little bit trickier to reset, though. Have been struggling with these in my current repair efforts. As a crappy zwave network is not an option, I guess I will just have to bite the bullet and try a fresh start following the steps you reference.
Those in wall switches, perhaps you can reset them with the wall switch? Triple on / off or something similar.
Yes, I finally manage to include a number of those as well. I am rebuilding my zwave network slowly now, starting with only powered entities saving the battery ones for later.
I have had a couple of hickups during my first 7 inclusions in the way that HA doesnât recognize any zwave entities (just showing empty in the zwave config). So far, these âblind spotsâ have been remedied by a "hassio host reboot".
Unfortunately, this confirms that zwave under HA is rather âwonkyâ at times.
I had to rebuild my zwave network over the weekend after I realised that a repeater had traded node identity with a switch. This lead to massive OZW_Log.txt files due to node INFO spam and some non-functioning automations ; my working theory was that this was the result of multiple restarts without allowing time for OZW to stop correctly.
I am having similar problems. Iâve installed a zwave aeotec z stick gen 5 and a four in one sensor. Get everything up and running, but it just doesnât work after a reboot. The stick shows as âunknownâ and the node shows âentity not availableâ. Happens each time I reboot HASS.io. This is not the first time either. Not sure what is going on, but I canât get it to work consistently or reliably.
Iâm running HASS.io in a docker.
Its just gone and done it againâŚfollowing a restart all my nodes were gone!!
This time Iâve resorted to doing Z-Wave network stops and restarts and network heal operations. I now have most nodes responding! I seem to have lost one door sensor that so far has refused to come back.
I am tired of this!!!
Neither - I am running Docker. I usually do a docker restart hass to restart the system
Make a backup copy of your ozwcfg.xml file, issuing a docker restart doesnât allow the OZW stack time to shutdown properly and causes it to mess up the xml file.
When firing up your docker backup copy the backup over the ozwcfg.xml before hand and you should be fine.
Would stopping HA from the Configuration > General and then bouncing the docker be the preferred method then? With as many breaking changes I have seen in the last few updates HA docker gets bounced quite a bit to sort them all out. Would be nice to know the âbest practicesâ for doing a restart. Earlier versions around 0.7x didnât seem to care but I have noticed since around 0.84 or so z-wave nodes will fail to come back. And the latest 0.95 without an entity registry file seems to like it even less. Had of the same devices showing up for hours today
Whenever I do any serial restarting of HA (restarting over and over) I always wait until all of my nodes show as âreadyâ for all of my mains powered devices. By then the battery powered devices should all be back up too.