1242AG AP's disassociating themselves with 4400 controller
I have a single 4400 WLC(4.0.2060)and seven 1242AG AP's running IOS 12.3 covering our warehouse. The problem is that a few times a day we get all or most of the AP's disassociate from the WLC without any warning.
Here are a few observations/thoughts on my part:
1. Everything had been running stably for months until a week ago when we added a bunch more handheld scanners for our warehouse to deal with our peak season, increasing total wireless clients from around 130 to about 180. I notice that the AP client load can get very uneven with some AP's getting 50 to 70 clients at once. As people pick inventory we can get clusters of activity that are in one area. I know that Cisco recommends around 25 clients per AP. Could this be the problem? Do I need more AP's?
2. I noticed that there are lot of undersized packets reported on the switch ports that WLC distribution ports 1 and 2 are plugged into. We disabled LAG and the undersized packets continued. Now I have seen bug note CSCse06509 and as we are using a 3750G switch I assume this is the problem. Will 802.3 flow control help?
3. I have given some thought to increasing the Arp timeout on the WLC. Any thoughts?
4, The WLC is not in master controller mode and the AP's have static addresses.
5. Spanning Tree Protocol (STP) is enabled in portfast mode on the switch ports and active all trunked VLAN's but disable on the WLC. Is this a problem?
6. I an using Layer 3 LWAPP. But all the AP's are on the same subnet as the WLC so I supposed I could switch to layer 2 if that might help.