This is a weird one. The environment looks like this: 5520 controller running 8.5.140. 3702i APs. Spectralink Versity 9553 or Apple iPhone SIP clients. SSID is local. QOS is set to Platinum.
Infrastructure looks like this:
5520 <- VPC over 10Gig Coax -> Nexus 7k DC switch <-VPC over 10G fiber -> Cat2960 Closet switch <-> 1G Copper <-> 3702i AP1
Phones have connectivity. Can ping, surf the web, etc. Phones can make outgoing calls fine. Incoming calls to phones do not ring. SIP traces just show "Trying" repeatedly. Phones never respond.
If I disable this AP1, the phone connects to a nearby AP2 (on a different switch) and processes incoming calls fine.
If I change QOS f on the WLAN from Platinum to Gold, incoming calls are also processed fine.
If I then go back to AP1 on Platinum QOS, incoming calls fail again.
I have gone through all of the "plumbing" - interfaces and port-channels between the DC and the two closet switches are very similar. Nothing looks out of place. I have swapped the AP out with a different 3702i and no change. This is the only location in my entire enterprise (>1000 APs) that I know of where inbound call processing fails like this. I do have a TAC case opened on this, but I figured more is better in terms of finding a solution (or at least some empathy).
Any ideas on how to proceed would be appreciated!
I would simply swap AP1 & AP2 , if the problem follow with AP1 then you know it is AP specific.
Once you do that, if the problem appears with AP2, then it could be that specific switchport. I would configure a new switchport and test if that is the case.