Technical data

N+1 Redundancy

A set of master controllers and a standby slave controller are configured via static IP

addressing to reside in the same subnet, and are considered to be an N+1 cluster. The

standby slave monitors the availability of the master controllers in the cluster by

receiving advertisement messages sent by the masters over a well known UDP port at

expected intervals. If five successive advertisements are not received, the standby

slave changes state to an active slave, assumes the IP address of the failed master,

and takes over operations for the failed master. Because the standby slave already

has a copy of the master’s latest saved configuration, all configured services

continue with a short pause while the slave switches from standby to active state.

While in the active slave role, the slave controller’s cluster monitoring activities are

put on hold until the failed master rejoins the cluster. An active Slave detects the

restart of a master through ARP. When the active slave is aware of the master’s return

(via the advertisement message) it relinquishes the master’s IP address and then

returns to the standby state. The now-passive slave will not fail over for the same

master until a WTR is completed.

If it is necessary for the failed master to be off-line for a lengthy interval, the admin-

istrator can manually set the active slave back to the standby slave, thereby ensuring

the standby slave is able to failover for another master.

In most cases with a cluster of N+1 Masters, the APs all have to be in L3 Connectivity

mode, but if you only have one Master and one Slave unit (N=1) the APs can be in L2

connectivity mode. In this case, while the Master unit is active the Slave unit will not

take AP registration so the AP will always go to the correct controller.