I’ll try to condense this as much as possible because there are a lot of moving parts.
This is purely a learning exercise for my homelab, so try to answer based on the tech I’m using and refrain from alternative tech unless it’s absolutely required.
Proxmox CE: Hypervisor OPNSense: Router between ISP and my network Kubernetes (Talos): qemu-agent-enabled ISO to work with Proxmox
I have 3 control planes: k8s-cp-1 through 3. I have 4 worker nodes: k8s-w-1 through 4. All of these are VMs in a single Proxmox instance.
k8s-cp-X: 192.168.0.51-53 k8s-w-X: 192.168.0.54-57
(All IPs are DHCP reserved by MAC and they are outside of (under) my automatic DHCP range of 100-254)
These instances are not joined to a cluster at the moment (they’re all in maintenance mode).
I’m trying to create a VIP in OPNSense at 192.168.0.50 that failover to my control planes so that I can forward my service ports to 192.168.0.50 and the ingress handle the rest based on hostname. I created the VIP (as IP Alias), but I’m lost as to how to do the failover part. In my mind, I’m thinking that pinging .50 should respond from whichever of the 3 control planes IPs is currently being routed to from the VIP, but I don’t know how to make that work.
Hopefully, that all makes sense. Oh, and I’m thinking Cilium as LB and Traefik as Ingress controller, but open to suggestions there. I have experience with Traefik and Docker Swarm which is why I was thinking of using it for Kubernetes.
