Hi guys.. here again with some questions.
I have this scenario with a WLC 4402 running 184.108.40.206 and 47 APs configured. I got this problem where randomly, some of the APs reset but there seems to be some kind of synchronization, because for example, 5 APs restart at the same time, being in different switches, being different models (1252 and 1242) using different types of power (PoE and pinjectors). The switches remain on (verified with the show ver command), so is not because the switches failed or something... ive tried to get crash info from the APs but for some reason I cant get it from the controller (with the config ap get-crash-file command), and if i telnet to the APs, the only info I can get is that they returned from a power on.
Any ideas on this? APs and controller are in the same segment.
How do you have the interfaces configured on the wlc? How many of the 1252's do you have and the 1242? Does it matter which ap or how it is powered? Tell us a bit more of how everything is setup....
The interfaces are configured as portchannel. We have 25 1242 and 22 1252. Almost all of the switches do PoE. Only 5 - 10 APs are powered using power injector.
The weird thing is that I cant find the similarities between the APs that restarted at the same time, some of them are in the same switch, some aren't, some are in different buildings. And weirder: not all of the APs in the same switch restarted at the same time. For example, in the same switch one 1242 restarted 50 days ago and the other 7 days ago.. at the exact same time as 8 more APs from different switches.
On the switch that the wlc is connected to run a:
show etherchannel load-balance
It should be set to src-dst-ip, if not then run on the switch:
port-channel load-balance src-dst-ip
Look at the ap's when they have joined a wlc and see how long the ap has been up, how long the ap was joined with that wlc.
As you said, "src-dst-ip" was configured.
The AP Uptime:
UP Time 24 d 0 h 32 m 45 s
Controller Associated Time 24 d 0 h 21 m 12 s
Controller Associated Latency 11 m 32 s
This counters are almost identical among the other APs, (those in different switches, same and different models).
Well when a LAP looses connection to a WLC, the Controller Associated Time resets. Here they show 24 days. When a LAP reboots, the the uptime resets. So is this ap one of the stable ones?
Thats the problem, it looks like the AP rebooted, but i dont know if it rebooted because a malfunction (maybe power) or because it lost connection to the WLC, and went trough all the discovery process ending in a reload (getting DHPC address again). But.. why so many APs at the same time? Of the 47 APs I got like 4 groups.. 10 with 50 days up.. 10 with 24 days up.. some with 10 days up and so on... but each group looks like they restarted at the same time.
Hi, I was just reading this and I wanted to ask if you have the 4402 in LAG mode, where both WLC ports are joined to make one big interface. I was recommended to do this when I added more than 25 AP's to one of my networks. I guess it's a potential bandwidth problem if you have more than 25 AP's on only one port...
Any relevant controller traplog messages around the time these disconnects occur? What do you have for capturing syslogs/traps?
Configure the APs to syslog to your desktop using the Kiwi Syslog Deamon. You may see some messages indicating what is going on.
What's the network latency between these APs and the controller? Where are the ones that have stayed up the longest? What PoE switches/linecards are you using?