wlc 4404 strange behaviour

caz-network · ‎12-27-2007

Hello,

I have to manage a wlc 4404. According to me he acts strange.

Only ports 1,2 and 3 are connected. The ap-manager and the management are linked at port 1. So far so good.

If I look at the wireless-tab I see all AP's are connected at port 3 !!

All the 3 ports are connected to the same switch and VLAN.

AP=Cisco Aironet 1130AG with a default config

Could someone explain me why all AP's are connected on port 3, while the ap-manager is linked at port 1.

Thanks in advance,

Carlo

Scott Fella · ‎12-27-2007

Are you using LAG on the WLC and is the etherchannel up on the switch side?

-Scott
*** Please rate helpful posts ***

dennischolmes · ‎12-27-2007

Sounds like he is not using LAG. LAG or etherchannel is a way of bonding the physical ports 1-4 to a single logical IP address. If you do not do this then you need to setup an AP manager IP address for each sfp port on the controller that you have active.

caz-network · ‎12-28-2007

I am not using LAG. I have setup an AP-manager on port 1 not on port 3. But all AP's are connected on port 3. At the moment I have 48 AP's.

I add 15 new AP's on the same switch, three of them can connect to port 1 (ap-manager), for all the other I see the following message :AP cannot join because the maximum number of APs on interface 3 is reached. Very strange.

Additional information: port 2 has a different interface, it also isn't an ap-manager, and it is connected to the same switch as port 1 and 3, but in another VLAN.

I can try the following:

I disconnect port 3, normally the first 48 AP's has to switch to port 1. I create a new ap-manager2, linked to port 3.

The new situation:

port1 managment + ap-manager

port2 a none-ap-manager interface

port3 ap-manager2

dennischolmes · ‎12-28-2007

Go to controller/interfaces/ap-manager/ and verify that the physical interface is port one and that it is on the same vlan as the APs. Quite often, a bug in the code resets the physical port to 0 (zero) when you reboot for the very first time. Other than that and a possible vlan misconfiguration on the switches, I'd call TAC.

caz-network · ‎12-28-2007

Physical interface is port 1.

After the hours I am going to disconnect port 3. I hope that all AP's will switch to port 1. If not I hope the logging will give me some light in the darkness.

Thanks for your input.

caz-network · ‎12-31-2007

Problem is solved !

I added a new ap-manager interface active on port 3. I disconnected port 3, 48 AP's switched to port 1. After I reconnected port 3 all other AP's connected to port 3.

rseiler · ‎12-31-2007

The Cisco docs clearly state that you can only have 48 APs associated to each ap-manager interface. If you are not running LAG, and I can't think of a good reason not to, then you will need 3 physical interfaces to be configured with unique ap-manager ip addresses in order to register 100 APs. Since the APs point to the management ip (via DNS or DHCP) to get the ap-manager interface, they will automatically load balance.

Note that unless you are running VERY recent code, there are significant ARP bugs related to the additional ap-manager interfaces responding to ARP requests. The best-practice is to add a static ARP entry in the router/L3 switch that is the ap-manager's default-gateway. This is an artifact of the ap-manager interfaces not responding to any traffic accept LWAPP, including ARP. This has been a serious problem for a lot of enterprise customers because this impacts most CEF switching devices like the Catalyst 6500 since most IOS versions also have a CEF adjacency/ARP timeout bug.

dennischolmes · ‎12-31-2007

I stated this in an earlier post. See the above.