7921G Phone Call Drops when switching WAPs on Different WLC's

[rholtby]

We are having issues with calls getting dropped when a 7921G phone traverses from a WAP that is connected to Wireless LAN Controller A and transfers to a WAP on Wireless LAN Controller B. This is occurring with both of our 4402WLC and 2106 WLC. We have mobility groups enabled and setup and they can ping the other controllers. There is no problems when traversing from one WAP to another WAP that are associated with the same WLC controller. Any ideas where the problem may be?? I'm trying to find information on configuration settings for this but not much information out there.

[Richard Atkin]

Can you tell us what authentication / encryption mechanisms have been deployed?

Thanks.

[rholtby]

WPA2 with pre-shared key. We haven't setup data or guest lans yet simply trying to get the voice network going at this point.

[migilles]

Would encourage to use 4.1.185.0 or 4.2.61.0.

But since no issues w/ intra-controller roaming, seems to be a controller config or bug or network issue.

[rholtby]

We saw the problem with 4.1.185.0 and have since upgraded to 4.2.61.0 and still have the problem with the lastest code.

[john.preves]

sorry, I have to ask the obvious...are these two sets of controllers in the same vlan? You are not attempting to roam between the 4xxx and the 2xxx on two vlans?

[john.preves · Accepted solution]

one more ...are the ssid's in the wlan groups between both controllers pointing to the same interface?

[rholtby]

The SSID's and the WLAN groups are identical on each side. In fact we deleted and recreated them a few times without any luck. The 4402 controllers show the physical interface as LAG and are connected via channel-group on a 4506 controller. The 2206 controllers are connected to a 3750 switch and the physical interfaces are pointing to interface 2. We have a channel-grou setup for each of the 2 gig interfaces on different 4506 controllers (Primary and secondary) to the 4402's. The 2206's connect through a 3750PS Switch and are connected to the same switch. When all the access points are forced to one controller we do not have any issues with handoffs between WAPs. Again it's when they are load balanced that we see this handoff problem.

[rholtby]

On your devices and interfaces verify

Core to Core Port Channels - (This is where you have one 4402 or 2106 on one core and one on another core switch)

interface Port-channel1

switchport

switchport trunk encapsulation dot1q

switchport trunk allowed vlan XX, XX -- Where XX is your voice, controller and Data Vlans etc..

switchport mode trunk

Core to WLC

interface Port-channel3

description LAG trunk to WLC01

switchport

switchport trunk encapsulation dot1q

switchport trunk allowed vlan XX, XX, XX -- Where XX is your voice, controller and Data Vlans etc..

switchport mode trunk

WLC Controller

Also Verify and ensure LAG mode is enabled on the WLC under:

Controller->General->Enable LAG Mode on next reboot

We also updated the phone firmware to the most current release but this did not actually solve the problem.

[rholtby]

The two controllers are on the same vlan.

[john.preves]

I'm pretty sure I didn't understand any of that..so it's probably me, but your phones have the same IP address on them when they were on controller A as they do when they roam to B?

I apologize in advance for my inexperience.

Other than the above, I am over my head.

[rholtby]

We are on the same subnet and we haven't setup inter subnet roaming. I haven't checked and this is a good question. We will double check on this. Cisco TAC is getting us to upgrade the phone firmware as it's down two levels. TAC support says all the roaming and traversing information is handled by the phones. We are at version 1.04 and we should be at 1.06. Our callmanager version 5.1.2. We are planning to upgrade the firmware on the phones later today.

[rholtby]

On your devices and interfaces verify

Core to Core Port Channels - (This is where you have one 4402 or 2106 on one core and one on another core switch)

interface Port-channel1

switchport

switchport trunk encapsulation dot1q

switchport trunk allowed vlan XX, XX -- Where XX is your voice, controller and Data Vlans etc..

switchport mode trunk

Core to WLC

interface Port-channel3

description LAG trunk to WLC01

switchport

switchport trunk encapsulation dot1q

switchport trunk allowed vlan XX, XX, XX -- Where XX is your voice, controller and Data Vlans etc..

switchport mode trunk

WLC Controller

Also Verify and ensure LAG mode is enabled on the WLC under:

Controller->General->Enable LAG Mode on next reboot

We also updated the phone firmware to the most current release but this did not actually solve the problem.

[fcoppotelli]

I had your same problem.After many tests I fixed the situation configuring all the AP's on one WLC and keeping the second one in standby (empty but ready to accept registration if the first fail).

You can configure the order of the WLC's on each AP.

Regards,

Fabio

[pfocas]

Hi

If the break is being caused by a re-authentication, it may be worth trying to use CCKM. This prevents a reauthentication on a roam, and is only supported using TKIP, not WPA2 (I presume you're using AES). I'd use either EAP-FAST or LEAP and CCKM to see if it makes a difference.