kubectl exec & proxy broken after installation of gke example

[ghost]

When running a regional cluster it seems that both kubectl exec and kubectl proxy are completely non-functional and broken.

My terraform.tfvars looks like the following.

region = "europe-west1"
zone = "europe-west1-b"
gke_master_ip = "35.xxx.xxx.xxx"
gke_node_tag = "gke-xxxx-xxx-node"

Any suggestions?

[danisla]

Looks like this is because regional clusters have multiple masters and the cluster Endpoint is actually a load balancer. By inspecting the firewall rule postfixed with -ssh you can find all of the master IPs that need special routes to bypass the NAT gateway.

I have a fix and it will be pushed soon!

Thanks

[ghost]

👍 Thanks a ton!