Product Documentation

Understanding the Causes of Failover

Sep 30, 2013
The following events can cause failover in an HA configuration:
  1. If the secondary node does not receive a heartbeat packet from the primary for a period of time that exceeds the dead interval set on the secondary. (See Note: 1.)
  2. The primary node experiences a hardware failure of its SSL card.
  3. The primary node does not receive any heartbeat packets on its network interfaces for three seconds.
  4. On the primary node, a network interface that is not part of a Failover Interface Set (FIS) or a Link Aggregation (LA) channel and has the HA Monitor (HAMON) enabled, fails. (See Note: 2.)
  5. On the primary node, all interfaces in an FIS fail. (See Note: 2.)
  6. On the primary node, an LA channel with HAMON enabled fails. (See Note: 2.)
  7. On the primary node, all interfaces fail (see Note: 2). In this case, failover occurs regardless of the HAMON configuration.
  8. On the primary node, all interfaces are manually disabled. In this case, failover occurs regardless of the HAMON configuration.
  9. You force a failover by issuing the force failover command on either node.
  10. A route monitor that is bound to the primary node goes DOWN.
    Note: 1. For more information about setting the dead interval, see Configuring the Communication Intervals. Possible causes for a node not receiving heartbeat packets from a peer node include:
    • A network configuration problem prevents heartbeats from traversing the network between the HA nodes.
    • The peer node experiences a hardware or software failure that causes it to freeze (hang), reboot, or otherwise stop processing and forwarding heartbeat packets.
    Note: 2. In this case, fail means that the interface was enabled but goes to the DOWN state, as can be seen from the show interface command or from the configuration utility. Possible causes for an enabled interface to be in the DOWN state are LINK DOWN and TXSTALL.