How Cluster Redundancy Works

The following simplified scenario describes how cluster redundancy works and how managed APs fail over from one controller cluster to another.

  • Active-Standby mode

    This mode offers limited UI configurations as most of them are read-only configurations on Standby cluster.

    1. After you enable and configure cluster redundancy on the controller, managed APs will obtain IPs of all nodes in Active cluster as server list, and all IPs of all nodes in Standby cluster as failover list, which is shown in AP as:

      {

      "Server List":[ "IP_A1", "IP_A2, "IP_A3", "IP_A4"],

      "Failover List":["IP_B1", "IP_B2, "IP_B3", "IP_B4"]

      }

    2. If Cluster A goes out of service or becomes unavailable, APs managed by Cluster A will attempt to connect to the IP addresses (one node at a time) specified for Cluster A.
    3. If managed APs are unable to connect to the IP addresses specified for Cluster A, they will attempt to connect to the IP addresses (one node at a time) specified for Cluster B.
    4. If managed APs are able to connect to one of the IP address specified for Cluster B, they fail over to Cluster B. APs will move to the zone it belongs to when failover.
    NOTE
    The standby cluster to which APs fail over must have sufficient license seats to accommodate the new APs that it will be managing. If Standby cluster has insufficient license seats, some APs may not get HA license and these APs will be rejected by the standby cluster.
  • Active-Active mode

    Configurations can be made using the UI.

    1. After you enable and configure cluster redundancy on the controller, the IPs of failover list come from all the target active clusters (up to 3) configured in current active cluster are prioritized per cluster, but the nodes in cluster are randomized.
      For example, if you enable the cluster redundancy with active-active mode on current active cluster A and configure following active clusters with priority:
      1. Cluster B
      2. Cluster C
      3. Cluster D

      The managed APs will obtain IPs of all nodes in cluster A as server list, and all IPs of all nodes in target active clusters as failover list, which is shown in AP as:

      {

      "Server List":[ "IP_A1", "IP_A2, "IP_A3", "IP_A4"],

      "Failover List":["IP_B4", "IP_B2", "IP_B3", "IP_B1"], ["IP_C1", "IP_C4", "IP_C2", "IP_C3"], ["IP_D2", "IP_D1", "IP_D4", "IP_D3"]

      }

    2. If Cluster A goes out of service or becomes unavailable, APs managed by Cluster A will attempt to connect to the IP addresses (one node at a time) specified for Cluster A.
    3. If managed APs are unable to connect to the IP addresses specified for Cluster A, they will attempt to connect to the IP addresses (one node at a time) specified for Cluster B, and will try next Cluster C if APs unable to connect the IP address (one node at a time)specified for Cluster B.
    4. If managed APs are unable to connect to the IP addresses specified for Cluster C, they will attempt to connect to the IP addresses (one node at a time) specified for Cluster D, and will start all over again from Cluster A if all IP addresses unable to connect.
Parent topic: Cluster Redundancy