Access Point Switchover
Switchover driven by Access Points
In this scenario, there is no direct communication between primary or secondary WLCs. But switchover of traffic is driven by Access Points.
Access Points are manually configured (via Management Console/GUI), with Primary and Secondary WLC information. AP automatically switches to secondary controller if primary is not responding.
When Primary WLC goes down, AP will stop receiving ‘keep-alive’ request from it.
By recognizing the loss of this request, AP starts sending join request to secondary WLC IP (which is already configured), and establishes the connection. Once primary is up, AP again reverts its connection toward primary WLC IP.
This is simplest failover configuration and the process may cause interruption of wireless service to user.
The WLCs are not HA aware and no heart beat happens between WLC’s. However, an AP will always be up and running under any one of the WLCs.
AP-SSO: Access Point Stateful Switchover (Cisco WLC High Availability)
The Cisco SSO technology helps mitigate and improve user experience by reducing the downtime faced by the user by utilizing a secondary standby controller connected to the primary controller via HA link over a dedicated redundancy port. In this case, when the primary controller goes down the AP can move over to the secondary controller instantly, this means the client devices will not experience any SSID downtime. The Cisco SSO technology also transfers the security keys of all the client devices connected through the AP, to the standby controller, so that the client can re-authenticate seamlessly.
Cisco WLC High Availability.supports 1:1 (Active:Standby-Hot) stateful switchover of access points (AP SSO).
Before HA is configured, both the controllers will have their physical IPs and management IPs separately. Once HA is established, both the controllers will share common management IP (which is of primary) and primary acts like active, and secondary in Standby-Hot mode.
There will be a heartbeat between WLCs and automatically if one WLC goes down the other will be active.
CA Spectrum 10.2 leverages the Cisco SSO technology where, the two controllers are always in sync with a duplicate copy of the AP information.
As per Cisco design, only active WLC is enabled with SNMP. In AP SSO, when the currently active controller (lets say Primary) goes down, the standby controller (Secondary) becomes active and SNMP enabled.
When active WLC is modeled in CA Spectrum, standby WLC will be automatically modeled as Pingable (since SNMP is disabled on standby, as per Cisco design).
If secondary WLC is already down, at this time an alarm will be raised on the primary model, to indicate that the secondary WLC is down.
When the primary controller goes down the secondary controller becomes active, and its model will be upgraded from pingable to SNMP.
The SNMP community string, auto-placement and enable/disable of AP discovery, secure domain information, etc, for secondary WLC model, is referenced from the primary WLC model.
Secondary WLC would be remodeled as an SNMP device model (i.e. with new model handle) replaces the earlier (pingable model) with an snmp model.
The Secondary (standby) controller will be upgraded to SNMP (active) based on two conditions:
- Primary controller contact status is down
- Secondary controller (pingable model) receives (Lost Peer, Moving to Active-No-Peer State) trap.
Due to a limitation in the Cisco SSO technology, SNMP is enabled only on the Active WLC. Hence, Management Agent Lost alarm is generated only on the secondary WLC when the switchover happens from secondary to primary WLC.