Wireless Access

I have a network with 8 Controllers, 4 controllers in each region (1 Master and 3 Locals).

Masters are working in Active/Standy.

For Locals, we have VRRP running between them for redundancy. For each region it is like this:

Local 2 is backup for Local 1

Local 3 is backup for Local 2

Local 1 is backup for Local 3

For each AP group, LMS IP is the virtual IP for that region while Backup LMS is VRRP virtual IP from another region (So an access point terminated on Region A-Local1 will have redundancy as: Region A Local 1 (VRRP Active for LMS) -> Region A Local 2(VRRP standy for LMS) -> Region B Local 1(VRRP Active for Backup LMS) -> Region B Local 2(VRRP standy for Backup LMS)).

We have all good upto this point.

My question is if I need to configure fast failover in this scenario, how exactly that would work. I can configure an HA group with Region A Local 1 and 2 as members. In case Local 1 goes down, AP will move to Local 2. For the next tier of redundancy (in case both Locals in Region A goes down and Access Points move to Region B VRRP VIP, i.e Backup LMS), it won't be fast failover?

Anyone??

It is hard to understand what you have without a diagram, but I will say:

- Keep redundancy within the same region.  It is very unlikely that two controllers will fail at the same time and if they do you are likely to have bigger issues than wireless.

- Keep redundancy within the same region.  Most people who extend redundancy out of the primary region inherit alot of administrative overhead trying to manage what controller what AP should be on.  In larger, more complicated setups like yours it is best to have APs only expected to be on one controller or the other, or you end up having to chase down APS across controllers during upgrades, downgrades and switch maintenance.

What you are doing inside your region should be good enough for everything besides the zombie apocalypse.

Others have tried to extend redundancy outside of the existing region and it has just cause more issues because everyone forgets the sequence of failure and cannot explain it to anyone else why they configured it that way when take the day off from work....

Its not about me actually. It is the customer who is asking for this level of redundancy.

They are one of the biggest ISP in middle east region and have a big Public Wi-Fi network. They had an existing Cisco based network that gave them Primary, Secondary and tertiary controller option. I am working for an Aruba partner & to get into play, we had to propose same level of redundancy to meet RFP requirements. 

They want their network to survive in case one DC completely goes down.

Just to add, the network is working fine and as anticipated. I was just curious how fast failover can be added in current setup.

Without knowing anything, you can have an LMS-IP that points to a VRRP between two controllers.  You can have a backup LMS-IP that then points to a controller at another site, or a VRRP at another site.  That would put 4 controllers in play for redundancy.