Configuring Fast Failover and Enabling Session Availability

Before you configure the fast failover feature, ensure the following:

  • The primary and secondary controllers are properly configured in availability mode. For more information, see Availability.
  • The pair of controllers in availability mode is formed by one of the following combinations:
    • C5110 and C5110
    • C5210 and C5210
    • C4110 and C4110
    • C5110 and C4110
    • V2110 and V2110 (Using the same V2110 profile, two V2110 Small, or two V2110 Medium, or two V2110 Large.)
    • C25 and C25
    • C35 and C35
  • Both the primary and secondary controllers are running the most recent Extreme Networks ExtremeWireless software.
  • A network connection exists between the two controllers.
  • The APs are operating in availability mode.
  • The deployment is designed in such a way that the service provided by the APs is not dependent on which controller the APs associate with. For example, the fast failover feature will not support the deployment in which the two controllers in availability mode are connected via a WAN link.
  • Both the primary and secondary controllers have equivalent upstream access to the servers on which they depend. For example, both the controllers must have access to the same RADIUS and DHCP (Dynamic Host Configuration Protocol) servers.
  • The users (client devices) that use DHCP must obtain their addresses from a DHCP Server that is external to the controller.
  • Time on all the network elements (both the controllers in availability pair, APs, DHCP and RADIUS servers etc.) is synchronized. For more information, see Configuring Network Time.
    Note

    Note

    The fast failover feature works optimally in fast networks (preferably switched networks).

    To configure Fast Failover and enable Session Availability:

  1. Log on to both the primary and secondary controllers.
  2. From the top menu of the primary controller, click Controller.
  3. In the left pane, click Administration > Availability.
    Click to expand in new window
    Graphics/setting_avail_pair.jpg
  4. Under Controller Availability Settings, select Paired.
  5. Select the Fast Failover checkbox.
  6. Type the appropriate value in the Detect link failure box.

    The Detect link failure field specifies the period within which the system detects link failure after the link has failed. For fast failover configuration, this parameter is tied closely to the Poll Timeout parameter on the AP Properties tab Advanced dialog. The Poll Timeout field specifies the period for which the wireless AP waits before re-attempting to establish a link when its polling to the primary controller fails.

    For the fast failover feature to work within 5 seconds, the Poll Timeout value should be 1.5 to 2 times the Detect link failure value. For example, if you have set the Detect link failure value to 2 seconds, the Poll Timeout value should be set to 3 or 4 seconds.

  7. In the Synchronization Option area, select Synchronize System Configuration.

    This is a global parameter that enables synchronization of VNS configuration components (topology, role, WLAN Service, VNS) on both controllers paired for availability and/or fast failover.

    For more information about synchronization, see Using the Sync Summary.

  8. Click Save.
  9. Set the APs‘ Poll Timeout value for fast failover.
    1. From the top menu of the primary controller, click AP.
    2. Select the checkbox for one or more APs.
    3. Click Actions > Multi Edit. The Multi Edit dialog displays.
      Click to expand in new window
      Graphics/ap_multi_edit_poll_timeout.png
    4. In the Poll Timeout field, enter the poll timeout value in seconds.
    5. Click Apply.
Note

Note

The fast failover configuration must be identical on both the primary and secondary controllers. Logs are generated if the configuration is not identical. For more information, see the ExtremeWireless Maintenance Guide.

After you have configured fast failover, you can verify session availability to preserve the user session during the failover.