Technical data

N+1 Redundancy
© 2012 Meru Networks, Inc. Implementing Redundancy 93
N+1 with Non-Revertive Mode
Previously, N+1 feature was incapable of providing an option to control the failback
operation. This means, an active slave relinquishes a master's role as soon as the
master becomes operational. The service downtime is doubled, when the fail over
and fall back happens without the administrator's control. To control the fail back
operation a Non-Revertive feature is introduced.
The Non-Revertive feature enables the administrator to choose when an active slave
can relinquish a master's role. Apparently, an active slave verifies if the master is
ready to take over the master's role or not. If the master is not ready, the failback
operation fails to start, and active slave keeps operating as a master. While the
active slave keeps operational, the master stays in the standby mode, waiting for the
active slave to relinquish the role.
Here, the Administrator minds the failback operation. To control the failback opera-
tion, a non-revertive mode must be enabled. The failback operation can be sched-
uled at a desired date and time. A Revert can also be scheduled exclusively on the
Active slave. Once this is performed, the Active slave reverts back to Passive at the
time scheduled by the administrator. In another way, Active to Passive slave revert
is performed forcibly without scheduling it on the Active Slave.
See “Managing the N+1 Installation” on page 100 to configure the non-revertive
mode and to schedule the revert.
Preparing the Network
The N+1 cluster must be configured within a set of guidelines to operate as described
in the previous section. While configuring your network for N+1 redundancy, the
following guidelines must be followed:
 In the N+1 cluster, the slave and master controllers must be the same model and
run the same version of System Director software. A check is performed by the
slave controller after each master controller is assigned to it to ensure the
hardware model and System Director version are identical; if a mismatch occurs,
the slave is not allowed to switch over for this master, and that status is noted in
the Status display for the Master Controller.
 All master and slave controllers must use static IP addressing to ensure
consistency and control of N+1 clustering. (DHCP addresses are not supported for
controllers participating in the N+1 cluster).
 Master and slave controllers must be on the same IP subnet.
 All APs in the network should be configured for Layer 3 connectivity with the
controller.
 Spanning tree should be disabled on the switch port to which the controllers are
connected. To disable spanning tree on the port, refer to your switch
configuration documentation.
Configuring the N+1 Clusters shows a simplified network diagram of a recommended
N+1 deployment.