This message was deleted.

If your nodes have >1 network interface, you need to tell Calico how to pick which one to use. Calico calls that "autodetection method". https://docs.tigera.io/calico/latest/networking/ipam/ip-autodetection#autodetection-methods

how is that set through the Rancher UI? and for a given node - the nodes are heterogeneous from this view point

how is that set through the Rancher UI?

No idea, I work for Tigera on Calico 🙂

and for a given node - the nodes are heterogeneous from this view point

This is a global setting, so you need to pick the method sensibly. Choosing by interface name is usually not a good idea. The methods include things like "the interface that can reach a particular IP" or "the IP that kubernetes has recorded for this node".

This was adding a new node to access a specific new type of system into an existing cluster, I am trying to update the deployment at the moment to do a can-reach from the docs you linked and seeing if it makes a difference, notably there wasn’t an issue in a flannel environment that was tested before on another cluster but on the main cluster, calico seems to have a thrown some strange behaviour