I had what seems like a similar issue… my symptoms were:
- Z-Wave devices sporadically work
- Z-Wave devices are in PING/REQUEST_NIF for hours after restarting the binding
- Z-Stick G5 color stops changing and stick stops responding
- Z-Stick seems to start working after some period of time and nodes work for a bit until the stick “freezes” again
Environment: Windows Server 2016, Aeon Z-Stick G5, about 60 z-wave nodes
I removed the binding, all the things, node files, etc… multiple times and re-added but with no success. The z-wave network stopped working after some amount of time.
I downloaded Zensys tools and went through each node. I found 1 node that, when anything was done with it, would “freeze” the z-stick. This was a node that was previously excluded but still showed up on the z-stick. Anytime I tried to remove the node or send a NOP, the stick would completely stop responding and stop changing colors. Zensys was able to successfully control all the other nodes though with basic command sets.
I was not able to remove this node via Zensys or by doing an exclude again at the node, etc…
I backed up the z-stick and restored to a new one and the same thing happened, so it seemed like something was wrong with the config on the stick.
I ended up completely resetting the new stick and excluding/including all my nodes again, then deleting/re-adding all the things and re-mapping the items. This finally fixed it.
The “corrupted” node was excluded months ago and OH seemed to be functioning fine. Issues didn’t start until I upgraded to the 2.5 release. Not sure about causation, but there is correlation. Restoring OH to 2.4 ended up still having the same issues though.
Lesson learned: Make regular backups of the z-stick. This was outside of the scope of my backup strategy… that will be changing.
I definitely lost some WAF points with this outage, especially with it being over our vacation time (at home), but she did see how much we have come to depend on the automation.