There isn’t likely to be any retrospective fixes for that.
What times? From observation, this business is sometimes associated with nightly heals, stopping that can prevent unexpected NULLs
OK, thanks Ross!
Is it likely to have fixes for that in the latest OH 3 release?
No, I don’t think it has something to do with the nightly heals. It occurs from time to time from sensor to sensor over the whole day. Actually 3 minutes before for node11. 1,5 hours before now for node15. 3 hours before that no updates. My “feeling” for these thing updates is, their occurence has something to do with the wakeup interval setting (in the Fibaro device’s zwave settings - in my case it is actually set to 21600, resulting in thing updates every 6 hours, 4 times in 24 hours). I remember it was set to a time more often to update, but I reduced it already to 4 times a day.
I just don’t know why these thing updates change my items every time. If something really needs to be reconfigured - ok, no problem. This would be true if I changed my things, items, channels or whatelse. But I never did that or want that to do. All my things are auto-discovered through PaperUI ONCE, so things are under full control of the openHAB system, assuming to not get any problems like these now with zwave. At least, this was my intention from the beginning - and with the exception of zwave this intention was correct. I never had any problems with any other device/thing, and I have over 100 of them in my system.
Can I stop the nightly heals for testing? Are there any disadvantages I should be aware of when turning off the heals?
I don’t see why not. It’s possible the issue surfaces next time the device checks in after a heal - could be many hours later. Not enough known.
So far as I know they are redundant until/unless you add/remove or relocate devices.
But has the binding and its device database changed since then. I’m convinced one way to provoke this is to have an existing Thing created according to “old” database, when the current dbase entry is different in some way.
Pick a sick device, note the Thing UID, delete it. Re-discover/create a new Thing and make sure to give it the same UID at creation, it will link to existing Items etc. (assuming no channels change name or type). This is purely internal openHAB exercise.