I have over 20 1200 series PoE Aironets on campus, using multiple vlans + WPA.
Starting about 6 weeks ago, random waps will just stop responding to ping, ssh, http, and stops serving clients--I have connected via serial, and I dont get anything... Once I powercyle though, it works fine for an inderterminate amount of time, at which case it goes back to not responding.
I have switch PoE ports on the PoE switch, switched out to using PoE injecters, used a different switch, etc.
I dont have any ideas on what is going on....
What version of code are you running on the 1200 APs?
You mentioned that this started about 6 weeks ago -- were there any changes to the environment around that time? (Software changes, usage changes, # of users/clients etc)
Since the console port also is not responsive in this state we likely won't be able to capture useful information once the problem occurs. If one of the APs is experiencing the issue more frequently, you could setup a workstation to just monitor and save all console output until a failure.
There were no changes that I am aware of.
System Software Filename:
|System Software Version:||12.3(8)JA|
Alright, that software is a bit old, but I didn't find any specific problems that would cause this behavior in a quick search.
Do multiple APs get in this state at the same time, or are they all separate events? I'm wondering if some sort of traffic flood on the network (broadcast storm etc) is eating up the CPU.
I did some sniffing on this vlan (mgmt vlan) which all it currently has is the 20-odd aironets, and about 5 switches. I am seeing a bit of stp traffic, maybe a couple stp packets / second, with the current root bridge cost of 8.
Does that seem abnormal?
I'm not exactly sure what the expected amount of STP traffic would be in your environment. However, we can investigate the impact traffic might be having on your APs by checking the CPU usage on some of the APs:
show process cpu
show process cpu history
It really would be ideal if we could setup a console connection and monitor one of the APs until it hangs -- perhaps some console output around the failure point will explain what is causing this.