topic H-REAP with Centralized controller failover in Wireless

H-REAP with Centralized controller failover

twegmans — Sun, 04 Jul 2021 03:44:31 GMT

We have a setup with many remote locations that all have 4400 WLC's. We have 5508 controllers at several datacenters that act as centralized backup controllers in the event that the onsite controller dies. The AP's at these sites are set up in H-REAP mode because some of the clients are statically addressed and can't get a new addresses when the ap's fall back to the 5508 WLC. So we have the networks the statically addressed clients on set to locally switched via H-Reap.

This design does not seem to be working as anticipated. When the AP's at the remote site fall back to the central 5508's and become "connected" again the clients on the locally switched networks at the remote site no longer are able to communicate. I think that I must be missing something about how a H-Reap locally switched network acts when the AP connects to a secondary or tertiary controller.

H-REAP with Centralized controller failover

George Stefanick — Thu, 08 Sep 2011 20:27:36 GMT

I feel your pain...

Here a few ideas.

1) Put all the WLCs client side interfaces in the same L2. In fact, this is best pratices. This helps client from Re-IP addressing.

2) Do HREAP local swithing and let the clients get their IP from the local office. Its sounds like you are doing CENTRAL switching.

H-REAP with Centralized controller failover

twegmans — Fri, 09 Sep 2011 14:13:10 GMT

For the networks that we are having problems with we are doing LOCAL switching. The other centrally switched networks work fine. We were doing dr some testing in one of our retail store locations where we wanting to make sure clients continued to communicate in the event of a WLC failure. So we basically had someone onsite power off the local WLC at the remote location. Then from corporate we tested to ensure we were still able to communicate with the remote clients and also tested the remote client could communicate back to corporate. As i said in my previous post the locally switched clients are all statically IP'ed. That is really the only reason we are using H-REAP. Because these clients don't have the ability to fail back to the corporate controller and re-IP themselves but getting a new address from the corporate DHCP server

With the local WLC turned off -

-All remote site AP's failed successfully to the central office controller.

-Centrally switched network clients failed back to the corporate controller and obtained ip addresses the from the corporate DHCP server successfully.

-Clients on the locally switched networks could not get back to corporate and corporate could not get to them.

-We did not test is clients on a locally switched network could communicate to each other on the same subnet.

With locally switched networks shouldn't they just continue working un-interrupted whether the locally controller is there or not? How does it effect the communication flow of locally switched clients when the local controller goes away?

H-REAP with Centralized controller failover

George Stefanick — Fri, 09 Sep 2011 17:04:59 GMT

I agree. If your clients are HREAP they dump their traffic on the switch port and a controller failover shouldnt impact your clients.

So when the local WLC fails over what happens .. You cant ping the static client across the network ?

Also, Cisco released M1 code that allows for static address. Blogged about.. give this a read ..

http://www.my80211.com/home/2011/5/16/wlc-configuring-dynamic-anchoring-for-clients-with-static-ip.html