Dear all,
I have a setup where there are local controllers on each site (13 sites) and a master controller in the datacenter. When the local controller goes down, the ap's will failover the centralized master. On both the local and master are the same vlan id's but with different ip subnets since they are on different layer 3 environments. To support connectivity to voice clients we have set increased AP's "Maximum Request Retries" so that the ap will eventually drop it's ssid (since the local controller is down), then connect to the master and starts broadcasting the SSID again. The client will then reconnect and request an IP address. This works fine and the customer is ok with the short service interruption.
Now when the local controller comes back online, after the preemption period the AP will switch back to the local controller. This process however is so fast (just as fast as the original / factory default failover from local to master) that the ssid does not get interrupted, the client will stay connected. Better said: the client does not notice the interruption. Problem is that the client does not request a new IP address from the local subnet on the local controller.
Is there a way to influence this behaviour just as the failover process? Or maybe have a delay before the ap starts broadcasting the SSID again? This would make sure that the client will re-associate and request a correct IP address.
The above mentioned situation was also found in the forums, however there was no mentioning of the failback to the original situation.
Looking forward to your thoughts