I have a customer that has a handful of sites that have bridged APs, running on 6.1.3.1. They have defined the VLAN on the VAP to be 999 and have also defined the Native VLAN within the AP system profile as 999. This has worked for quite some time at those sites among a few different controllers. A couple of week's ago they had a power outage which brought one of the controllers down. When it came back up, one AP would only come up as "D" or dirty. When it was in this state, you could not provision it, it would not resopnd to apboot commands, nor could the technical support information be retrieved (it would hang the controller UI and CLI for 30+ minutes before time out). After troubleshooting for a while, we discovered that (despite working prior to that shutdown) if we change the Native VLAN in the AP system profile to 1, the dirty flag goes away. If we change it back to 999 it goes Dirty again immediately.
For now, we've left both teh VAP and AP System Profile Native VLANs defined as VLAN 1 as their is no real need to define the Native VLAN as the bridged APs are not trunked. This seems to be working fine. My question however is why changing it to 999 would cause a problem when it doesn't seem to at other sites and other controller (I did not try any other random number to see if it had the same problem).
Also, when this happened, the following was logged on the controller:
authmgr[1576]: <124063> <ERRS> |authmgr| Message to ip.of.AP:8223(SAPM Client) with MsgCode 0, Msglen 5, and Msgtype 3053 failed with Errno 145, Errstr Connection timed out
Any thoughts are appreciated.