This message was deleted.

dns settings are configured in the pod…

Mainly interested in setting the DNS search domain for the nodes and configuring the api to prefer host names for communication between the nodes instead of IPs.

RKE2 is configured to prefer IPs. You need static IPs for servers anyway.

Ok, maybe I’ll stick with IPs then. Sometimes we add a public NIC to worker nodes as the primary NIC which is on a public network that the control node is not on so we’ve had issues in the past where k8s was defaulting to using the IP of the first NIC when advertising the node address to the control plane which the control plane can’t reach in that scenario. So if we use a our_domain.local DNS search path and configure k8s to use hostnames that forces using the internal IP that we want whether it’s the first NIC or not. I know how to configure the flannel interface such as net1 (internal IP we want) instead of the default (net0 public IP in our case) which seems to work but I’m still seeing k8s use the public IP for node-ip so I’m assuming I need additional config aside from specifying the flannel interface? I need to avoid hard-coding the IP address but I can hard code the name of the interface…