What is a heartbeat; which links serve this role in a TR7 cluster?
Short answer
A heartbeat is the cluster members monitoring, with regular messages, that each other is up; if the messages do not arrive for a certain time, the member considers its peer failed and acts accordingly. In a TR7 cluster, this job runs over two separate paths. The first is the Sync Interface: the cluster connection between the two devices and the configuration synchronization are established over this interface; the interface is selected separately on each device with the Sync interface selector on the Hardware Management screen, and its status appears on the HA Clustering & Sync screen in the Cluster Status (Cluster connection is OK) and Sync Interface (Sync. interface (…) is up) rows. The second is for VIP failover: VRRP advertisements are sent between the Interface IP pair selected in the VIP Communication field on the interface where the VIP is; this path decides on which device the VIP will run. The Sync link going down does not move the VIP on its own.
Where It Is Used in TR7
| Where | Field / label | What for |
|---|---|---|
| Settings → Device Management → Hardware Management → Edit ON | Sync Interface selector (default None); Sync interface: 1 in the Usage cell of the selected interface | Selecting the interface that the cluster connection and the synchronization will use (device-specific, not synchronized) |
| Settings → Device Management → HA Clustering & Sync | Cluster Status, Sync Interface row (is up, has no link, is down, is absent in the system, is not set), Sync. Summary | The status of the cluster connection and the synchronization |
| Network → Interfaces → Edit Interface (in a cluster) | VIP Communication, VIP Communication (v6) | The path of the VRRP advertisements |
| Settings → General Settings → Notification Management | Notification Type: Cluster connection, Cluster synchronization, VIP connection status, VIP status | Notification when the cluster connection goes down, when the peer address in the VIP communication pair cannot be reached or when the VIP changes place |
Two Paths
| Sync interface | VIP communication pair | |
|---|---|---|
| What it carries | Cluster connection, configuration and database synchronization | VRRP advertisements |
| Where it is selected | Hardware Management (on each device) | The Edit interface form of the interface where the VIP is |
| IP address | Not added; no IP or route can be added to the sync interface | An Interface IP pair with an address defined on both devices is required |
| If cut | The cluster connection drops (Cluster connection is down); synchronization stops | The advertisements stop; the Slave device takes over the VIP |
TR7 also probes the peer address in the VIP communication pair with ping; when the probe fails, the VIP communication failed for <interface> message is generated in the VIP connection status notification. Details of the two mechanisms of the cluster: How does a cluster work?
Common Misconceptions
- "If the Sync link goes down, the VIP moves immediately." VIP failover depends on the VRRP advertisements on the VIP communication pair and the selected VIP Transition Method; the Sync interface is a separate path.
- "The Sync interface can be shared with production traffic." A separate interface is dedicated to the Sync interface; no IP or route can be added to this interface, it cannot be turned off, and it cannot be made a member of a Bond, LACP or Bridge (
Interface '…' is protected). - "The two paths are the same link." The Sync interface is for the pairing of the devices; VIP communication, on the other hand, is the Interface IP pair selected separately on each interface where a VIP is.
How it is done in the documentation
