.
Originally I just updated. After finding the problems with the healing, I have deleted all zwave things (via PaperUI), also the controller, and have added these again using the previous ThingID.
Will try the same using a new ThingID.
I now have deleted all zwave things (via PaperUI), also the controller thing, and have added these again using a new ThingID. Same result. Healing of nodes 8, 9, 10 ("non listening nodes) FAILED. The result is even worse. No manual healing is possible anymore. I could post the DEBUG.log here if anyone cares. If not, no matter.
You canāt run a heal on a node unless it has been fully initialized? Are you sure it completed initialization?
There are new xml files in zwave folder. All information inside, but ālast heal timeā.
All channels of the things are getting new values (motion, tamper, temperature, illuminance, battery), all OK.
Where else should I look?
Not sure. What happens when you select Heal the device in HABmin?
Set to heal in HABmin, wake up node manually.
node10:
.
But this morning I was able to heal node9. But I have no DEBUG.log
.
The other ones (node8 and node10) I didnāt test, because I need a ladder to wake them upā¦
Btw, I found out the following:
If a device is initialised, then you can see the following in āAdvanced Settingsā:
Then you are able to āReinitialise the deviceā.
.
If a device is not initialised, then you can see the following in āAdvanced Settingsā:
All my devices were āinitialisedā before I tried a heal.
Itās hard to say whatās going on with node 10. If I understand correctly, the heal fails when the device wakes up on its own, but succeeds when you wake the device up manually. My only thought is that the device stays awake longer when manually woken up versus when it wakes up on itās own. That would be a question for the manufacturer of the device.
1 Like
I donāt think so, that the awake period is longer, but you could wake it up very often in a row (by pressing many times the button).
What I now did: āReinitialise the deviceā for all three (8,9,10) nodes. The neighbors should be reseted. Now node10, has no longer any neighbors shown in āHABmin --> propertiesā. (Have to wait for node 8 and 9 awake period to see if this is also the case there). Hopefully during next heal the neighbors were recalculated. WDYT?
If your network is stable, why donāt you just disable the nightly heal and be done with it?
1 Like
Good question!
Just want to find out if I really have an error with the healing or if itās due to the devices.
Iāve now all reverted back to Snapshot #1486 (from 02 Jan 2019) and the heal is running again without any problems.
So it is now proven that there is a problem with the binding or something else e.g. OH-core or whatever.
Many thanks to all who have endured so long and supported me. 
1 Like
Have you tried upgrading just the binding in S1486 to see if the issue comes back? Or using the S1486 version of the binding in a recent OH snapshot? These tests might help to isolate the issue further.
No, I didnāt test it. Maybe Iāll find some time on All Saintsā Dayā¦ (A holiday in most parts of Germany.)
yesā¦ good thinking, then we can say if binding or core
@Celaeno1 go Alex go 
Let him go Mark, he nailed down the problem with the REST docs, heās on a role! 
1 Like
Why I went back has many reasons. There will soon be a new topic that is likely to cause a huge stir.
Here are the results of my testing:
Today Iāve tested: Snapshot #1731, with a downgraded z-wave binding #1486:
279 ā Active ā 80 ā 2.5.0.201901020757 ā ZWave Binding
Heal was running completly for: node9, node10, but it took 2 wakeup intervals!
node9:
node10:
.
node8: It was a failure while āassigning return routeā, but heal was done.
.
.
Iāve triggered a further heal, this will be complete in about 15 min. Iāll report itā¦
node9 and node10 were healed again!
node8: same result, failure while āassigning return routeā, but heal was done.
.
What does this mean?
1 Like
Alex, this is a brand new snapshot OpenHAB version with an old zwave binding from Jan 19 correct?
Yes, of course. See 3-4 posts above. Scott requested this testing! And you also found this a good idea.
1 Like