I have a GKE cluster running with the nginx-ingress
controller from the Kubernetes community (via Helm). I set it up with a regional IP which I had reserved in advance like so:
helm install stable/nginx-ingress \
--set controller.service.loadBalancerIP=my.reserved.regional.ip
I then added a couple of Ingress resources with the kubernetes.io/ingress.class: "nginx"
annotation and nothing much else, and pointed them to a headless service in the cluster (a type: ClusterIP
).
This works fine, and I can access the vhost-based ingresses via the load balancer IP.
However, I am wondering why the Ingress resources are also assigned an external IP:
==> v1beta1/Ingress
NAME HOSTS ADDRESS PORTS AGE
ingress-admin admin.my-domain.com 35.195.255.71 80 24m
ingress-api api.my-domain.com 35.195.255.71 80 24m
Is this normal? The IP responds to ping but does not serve HTTP (which my service does). Is there anything I need to do with the Ingress resources to avoid this?