HA with more than 2 nodes
We are currently expanding our hosting infrastructure, using OVH as a hardware provider and then do our own virtualization on the hardware.
All hardware has access to vRack and we have servers across multiple locations.
Currently our infrastructure is a bit vulnerable, as IP blocks are managed against each physical server, so 126.96.36.199/27 is on serverA, 188.8.131.52/27 is on serverB and so on.
We run pfSense virtual appliances, one for each physical server. But some VMs on serverA, might use an IP on serverB and therefore have the pfSense appliance on serverB as their gateway. This means that if the pfSense appliance on serverB og serverB goes down, these VMs would also be unable to communicate outside their local network/VLAN.
To resolve this, we were looking at making a pfSense appliance on each of our physical servers, configure them in HA and put our IP blocks on the OVH vRack, which would in return trade a bit of latency for redundancy, so that all traffic is always served from the master on serverA and if serverA goes down, serverB, serverC or serverD would then take over and only services running on serverA would be rendered unavailable.
But I cannot find any documentation on making a HA configuration with more than 2 nodes, so I was wondering what others are doing. Is there an option to do the HA directly between pfSense nodes or would we in these cases need to handle HA at the hypervisor-level?