Both appliances need to have the exact same time, it is recommended to use a local timeserver (e.g. domain controller, or dedicated timeserver)
The eth1 network interface of both appliances needs to be configured for the heartbeat accordingly
(e.g. 192.168.250.1 for primary and 192.168.250.2 for secondary node with 255.255.255.0 as network mask)
During the installation of the cluster, internet access is required for both appliances.
The system time must be equal on both appliances.
Allow the IP addresses of both appliances on your firewall for outbound mail traffic
The following steps are required to make a Cluster:
Select "Make Cluster"
Enter the IP and sf-admin password for the secondary appliance
Enter a Failover IP Address
The failover IP address is the common IP address the cluster is connected from the internal network e.g. firewall and mail server
Confirm the settings with "Make Cluster" and accept the following security question with "yes"
To fallover from one Cluster node to the other, select "Takeover":
This is useful if you need to put control to the passive node (e.g. if a hardware maintenance should be done on the active node)
The current active node turns to passive, the current passive one to the active node
To migrate a cluster, you have to be logged in on the webinterface via the failover IP
To release a Cluster, select "Leave Cluster" in the configuration, pay attention to the following Points:
You can only leave a cluster if logged in to the webinterface via the passive or active node IP, it is not possible to leave the cluster via the Failover IP
After releasing the cluster both appliances have the same data set.
Therefore only one appliance should be used for continuing operation, because otherwise emails can get sent twice.
That appliance you want to continue with must be rebooted.
The other appliance should be powered off.
Consider to reset this appliance to default settings before shut down
Notice that the network settings of the appliance have to be reset to your network requirements, so that the firewall and the mail server can connect to it.
To shutdown or restart a Cluster Node, select Actions -> Shutdown or Restart for the Cluster node
This function should be used with extreme caution, as it could damage the cluster
To fix a split brain state, select the node survivor (the cluster node with the most recent data) and select Actions -> Repair Split Brain
If the split brain cannot be recovered this way, release the cluster and recreate the cluster from the node survivor via "Make Cluster"