What's the restart count on the pod?
what's the output of
$ kubectl logs --namespace=kube-system exec l7-lb-controller-v0.5.2-apv4a -c l7-lb-controller
If it's restarting, can you please try:
$ kubectl --namespace=kube-system exec l7-lb-controller-v0.5.2-apv4a -c l7-lb-controller -- curl --silent localhost:8080/healthz
It should return a 200 and the string "ok".
If it doesn't, can you try:
kubectl --namespace=kube-system exec l7-lb-controller-v0.5.2-apv4a -c l7-lb-controller -- curl --silent http://metadata.google.internal/computeMetadata/v1/ -H 'Metadata-Flavor: Google'
And if that doesn't work, can you give me:
kubectl --namespace=kube-system exec l7-lb-controller-v0.5.2-0ldk1 -c l7-lb-controller -- cat /etc/resolv.conf
You should see the ip of your DNS service in there (kubectl --namespace=kube-system get svc kube-dns
). If it's there, and the nslookup/curl fails, can you exec into the kube-dns pod and try resolving the metadata service?
kubectl exec -it kube-dns-pod-name -c skydns -- nslookup metadata.google.internal
Your nodes need compute/rw scope, you can check this with:
$ gcloud container clusters describe foo
...
oauthScopes:
- https://www.googleapis.com/auth/compute
You should also be able to ssh into a node and try:
node $ gcloud compute backend-services list
Hi,
Here are my results: