41
Failover Events
In the occurrence of a Failover event the TallyMan system will move the Active status from the Primary
system controller to the Backup system controller.
Both units of a redundant pair monitor themselves and their partner for health. If a unit is deemed
unhealthy, then the redundancy will not be able to function correctly. For a unit to be deemed healthy
and therefore able to be part of a redundant pair it must meet the following criteria:
• The units are powered and fully booted
• Have stable network connections on the same LAN for each NIC. The second NIC will depend on
network configuration; if a configuration only has one NIC specified the second NIC will be
ignored.
• Both Backup Links, the Direct Link, and IP Link, are connected
• The units have a valid configuration that includes a redundant setup.
• The unit is in a ‘Running’ state meaning the internal system processes and services are running
and communicating correctly.
A failure on any one of these points will result in a system failover. The table below lists possible failover
scenarios:
During a Failover event various actions take place within the system to ensure the Primary is correctly
removed from service, the Virtual IP addresses have been removed and it is safe to apply them to the
Backup.
There are three resulting scenarios that depend on the failover event, the state of the Primary and the
network:
• Normal
• Error
• Remote Shutdown
A network cable used by the system controller is either removed from the switch or system controller, the
cable is severed or partially severed, or the network abruptly becomes either unavailable or partially
unavailable (e.g. no return comms).
The TallyMan software becomes unresponsive on the Primary unit,
Power is lost from the Primary system controller.
The network switch becomes unavailable either due to power loss or failure.
The system controller is overloaded with processing functions and is temporarily unresponsive.
A system controller’s network segment is disconnected from the network segments hosting the other system
controller either virtually or physically.
The Primary system controller experiences a hardware error that prevents it from functioning correctly.
Delay
There is a delay in communications between the redundant pair.