6.2.2. Procedure – Redundant heartbeat interfaces

Purpose: 

To avoid unnecessary takeovers and to minimize the chance of split-brain situations, you can configure additional heartbeat interfaces in PSM. These interfaces are used only to detect that the other node is still available, they are not used to synchronize data between the nodes (only heartbeat messages are transferred). For example, if the main HA interface breaks down, or is accidentally unplugged and the nodes can still access each other on the redundant HA interface, no takeover occurs, but no data is synchronized to the slave node until the main HA link is restored. Similarly, if connection on the redundant heartbeat interface is lost, but the main HA connection is available, no takeover occurs.

If a redundant heartbeat interface is configured, its status is displayed in the Basic Settings > High Availability > Redundant Heartbeat status field, and also in the HA > Redundant field of the System monitor. For a description of each possible status, see Section 23.7.1, Understanding PSM cluster statuses.

The redundant heartbeat interface is a virtual interface with a virtual MAC address that uses an existing interface of PSM. The MAC address of the virtual redundant heartbeat interface is displayed as HA MAC. The MAC address of the redundant heartbeat interface is generated in a way that it cannot interfere with the MAC addresses of physical interfaces. Similarly, the HA traffic on the redundant heartbeat interface cannot interfere with any other traffic on the interface used.

If the nodes lose connection on the main HA interface, and after a time the connection is lost on the redundant heartbeat interfaces as well, the slave node becomes active. However, as the master node was active for a time when no data synchronization was possible between the nodes, this results in a split-brain situation which must be resolved before the HA functionality can be restored. For details, see Procedure 23.7.3, Recovering from a split brain situation.

Note

Even if redundant HA links are configured, if the dedicated HA link fails, the slave node will not be visible on the High Availability page anymore.

To configure a redundant heartbeat interface, complete the following steps.

Steps: 

  1. Navigate to Basic Settings > High Availability > Interfaces for Heartbeat.

  2. Select the interface you want to use as redundant heartbeat interface (for example Physical interface 1). Using an interface as a redundant heartbeat interface does not affect the original traffic of the interface.

    Figure 6.5. Basic Settings > High Availability — Configuring redundant heartbeat interfaces

    Basic Settings > High Availability — Configuring redundant heartbeat interfaces
  3. Enter an IP address into the This node > Interface IP field of the selected interface. Note the following:

    • The two nodes must have different Interface IP.

    • If you do not use next hop monitoring on the redundant interface, you can use any Interface IP (even if otherwise it does not exist on that network).

    • If you use next hop monitoring on the redundant interface, the Interface IP address must be a real IP address that is visible from the other node.

    • If you use next hop monitoring on the redundant interface, the Interface IP must be accessible from the next-hop address, and vice-versa. For details on next hop monitoring, see Procedure 6.2.3, Next-hop router monitoring.

    Use an IPv4 address.

  4. If the two nodes are in a different subnetwork, enter the IP address of the local gateway into the This node > Gateway IP field. The Interface IP address of the node must be accessible from the Gateway IP address.

    Use an IPv4 address.

  5. Enter an IP address into the Other node > Interface IP field of the selected interface. Note the following:

    • The two nodes must have different Interface IP.

    • If you do not use next hop monitoring on the redundant interface, you can use any Interface IP (even if otherwise it does not exist on that network).

    • If you use next hop monitoring on the redundant interface, the Interface IP address must be a real IP address that is visible from the other node.

    • If you use next hop monitoring on the redundant interface, the Interface IP must be accessible from the next-hop address, and vice-versa. For details on next hop monitoring, see Procedure 6.2.3, Next-hop router monitoring.

    Use an IPv4 address.

  6. If the two nodes are in a different subnetwork, enter the IP address of the local gateway into the Other node > Gateway IP field. The Interface IP address of the node must be accessible from the Gateway IP address.

    Use an IPv4 address.

  7. Repeat the previous steps to add additional redundant heartbeat interfaces if needed.

  8. Click .

  9. Restart the nodes for the changes to take effect: click Reboot Cluster.