Within a failover cluster, the active node additionally takes over the Failover IP address on its network interface card
If, due to a malfunction, the active node fails, the secondary node will take over the failover IP address, thereby turning into the active node and staying accessible for the other network components, e.g. firewall and mail server, under the same IP address
It is not necessary to reconfigure the IP address
The heartbeat network gets installed using the two secondary LAN interfaces (LAN 2) of the appliances and a cross-over patch cable
Both appliances are controlling with the help of a regular impulse (heartbeat) whether the other appliance is still working properly
If the primary appliance does not react anymore, the secondary appliance will takeover all data resources and start the required services (engine and data base)
In case of a failover or an appliance breakdown the administrator will be advised by an email
In companies, there is an increasing desire in failsafe systems.
This can be assured with the REDDOXX Appliance in Cluster Mode.
To have a useful solution, the architecture and environments have to be considered.
We recommend the Cluster Mode in REDDOXX Appliances, when both Appliances are run as hardware.
A mixed operation between virtual machine and hardware, or even vm + vm is not recommended.
In order to have a stable running cluster, a direct heartbeat connection without a switch in between is required.
This can not be realised in virtual environments and might result in cluster nodes to enter a split brain state, where both nodes think that the other node is offline.
A split brain state can occur, when a short outage of the hearbeat net takes place (or not reachable due to dropped packages) and results in no or a delayed answer in the hearbeat net.
The other cluster node would then assume, that the connection is broken and the cluster partner has a problem.
This can result in serious consequential errors that are counterproductive to the main usage of a cluster.