Whether you contact support or not will be up to you. Unfortunately I suspect, at this point the only thing they can help with is removing ghosts, and you’re already doing that the likely answer you will get is ‘rejoin device X’ or ‘reset your hub’ …Nobody wants that.
So first, a little expectations setting. This will not be an easy, fun -or FAST process. You will be fighting this for DAYS.
(pause for effect)
BEST case, with the amount of weirdness you’re describing, expect you will be fighting this through the end of the week.
When removing ghosts and resetting routes, each repair will unblock something - but there may be other blockages and busted routes. You will do this again and again - clearing the next ghost you discover and removing it, then running a repair, until suddenly the backlog will clear and everything will suddenly work better. I was recently fighting with an issue post .34 firmware and it took me a record 22 repairs to get through it. (Yes, I counted. It took 14 days, and the entire time I was LESS than happy about it. I also have 80 something Zwave devices and it takes ~30 minutes for each run…)
It doesn’t matter if you have neighbors, doesn’t matter if you never added or removed anything, the unfortunate fact of the matter is your routes are busted and you have to fix them. Don’t go by whether the IDE reports a route or not - It’s only reporting the last reported route. With Zwave a given device can and should route automatically through all reachable neighbors - there will be be other routes in use by a device that don’t get reported.
Your bellweathers are ‘does my device work’ and does my repair come back clean. Of those two - all devices working ALWAYS trumps did the repair come back clean.
The best hints I can offer in fighting through my own:
- ALWAYS clear all visible ghosts before running any repair (I saw you have a few you didn’t delete - they’re not helping anything staying around - kill them)
- Resist every urge you have to spam the repair function. Repeated repairs don’t help much and usually needlessly complicate things. It takes time for the system to stabilize after running a repair. Best case is to let the system rest a day or so between repairs. Understandable sometimes you cant - but if you can find a stopping point, do it and let it ‘cook’ to see if things get better. ZWave devices will work behind the scenes on normal network timings and things and fix themselves - if given enough time to do so. I had better results running repairs if I waited between runs.
- If all of your devices save 1-2 are working - then kill and rejoin those few devices it will save a lot of headache.
- Locks are a pain in the BEST of circumstances because of the security requirements to joining. They’ll probably be the last ones to lock in and work. (My fiasco started with my back door lock)
- have a designated test plan to run after each repair (I had a spreadsheet with all the devices I cared about and tested every one in the list after each run.) If your list comes back good - stop, no matter what the IDE says. If it comes back good minus 1-2 devices rejoin those and then stop and let the network cook. (see #3)