default ipv6 subnet to /56

we were defaulting IPv6 subnet to a /64, this was working because kubeadm was calculating and modifying directly the node-cidr-mask, so the controller manager does not complain. Kubeadm should only validate that the config is correct and fail in case of error, this behavior will be changed soon, but once this happen it will break kind default deployment.
Huihuiyule-www · Oct 20, 2020 · 60074a9 · 60074a9
1 parent c9b98b3
commit 60074a9
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 2 deletions.
diff --git a/pkg/apis/config/v1alpha4/default.go b/pkg/apis/config/v1alpha4/default.go
@@ -51,7 +51,9 @@ func SetDefaultsCluster(obj *Cluster) {
 	if obj.Networking.PodSubnet == "" {
 		obj.Networking.PodSubnet = "10.244.0.0/16"
 		if obj.Networking.IPFamily == "ipv6" {
-			obj.Networking.PodSubnet = "fd00:10:244::/64"
+			// node-mask cidr default is /64 so we need a larger subnet, we use /56 following best practices
+			// xref: https://www.ripe.net/publications/docs/ripe-690#4--size-of-end-user-prefix-assignment---48---56-or-something-else-
+			obj.Networking.PodSubnet = "fd00:10:244::/56"
 		}
 	}
 	// default the service CIDR using a different subnet than kubeadm default

diff --git a/pkg/internal/apis/config/default.go b/pkg/internal/apis/config/default.go