Wireless Access

Hi,

We are looking at moving our current 6.5 deployment to a version 8 cluster in the near future and I want to clarify how the IP addressing is to be configured for AP termination. As far as I can see, in the user guide it states that managed devices are to communicate with the mobility master using their interface which has been designated as controller-ip. For AP's terminating on the cluster it seems to be a bit vague but I may be missing something - it looks to be that AP's should also terminate on the controller-ip (albeit via a VRRP interface) and this would be the same as is required using HA in version 6.x

In our current configuration, the WLAN controllers management network is separate to the access points  and access points cannot access the controller-ip interface. We would want to keep the management network separate in a version 8 cluster and would ideally keep all of the subnets we currently use for AP's the same.

Please could someone confirm if it actually is the case that access points must terminate on the interface designated as controller-ip when using clustering?

Hi, can anyone confirm my question? This is from the user guide:

The switch IP of the managed device is used as the IP address in the following configuration. The APs termination

point should also be set to the switch IP of the managed device. The LMS-IP for the AP in the AP system profile

becomes the active-AAC (A-AAC) for the AP.

(host) [cluster] (Classic Controller Cluster Profile "6NodeCluster ") #controller

10.16.116.3 group 1

(host) [cluster] (Classic Controller Cluster Profile "6NodeCluster ") #controller

10.15.116.4 group 1

(host) [cluster] (Classic Controller Cluster Profile "6NodeCluster ") #controller

10.15.116.5 group 2

(host) [cluster] (Classic Controller Cluster Profile "6NodeCluster ") #controller

10.15.116.8 group 2

(host) [cluster] (Classic Controller Cluster Profile "6NodeCluster ") #controller

10.15.116.9 group 2

(host) [cluster] (Classic Controller Cluster Profile "6NodeCluster ") #controller

10.15.116.10 group 1

(host) [cluster] (Classic Controller Cluster Profile "6NodeCluster ") #controller-v6

10.15.116.9

(host) [cluster] (Classic Controller Cluster Profile "6NodeCluster ") #controller-v6

10.15.116.10

APs will terminate to the controller-ip of the appropriate controller in the cluster.

What's different with AOS8 vs previous versions of AOS is that the AP will also receive the nodelist with the controller-ips of each controller in the cluster. With the nodelist, the AP is able to handle redundancy across more than two controllers, where-as the previous LMS-IP and Backup-LMS were limited to just two.

Thanks for confirming Charlie. Currently our AP's are split across 2 separate VLAN's and the controllers have an interface in each of these. We also have a management VLAN and the controller interface in this VLAN is designated as controller-ip. The AP's cannot access this interface. So it looks like I have some more re-design to do before we can start to use clustering