I am trying to understand how you can avoid using an external load balancer in kubernetes, if you want your service to be reachable on pods that live on different nodes. You would expect some virtual ip that points to your service so it gets routed to one of your pods. I have not seen an option (yet) how to configure this in Kubernetes.
So far I only have seen some other software “keepalived” that you need to install on all of your nodes to get the client request to your nodes.
An example of this is when you are using NodePort where all nodes listen to the same port and you point your client to one of the nodes. If the node, where the client points to, crashes, the client needs to point to another node in the cluster. By using a vip the client doesn’t need to know the IP-address of each node.
Do you have to use “keepalived” or can Kubernetes do this as well?