kubectl exec fail with 500 (dial tcp: lookup no such host)

[Egor Guz]

we register hosts with nova name and this name doesn't register in dns

kubectl get nodes
NAME LABELS STATUS
k8-pdbzkoijyr-0-ppo4rdunfyhl-kube-minion-agodggladxws kubernetes.io/hostname=k8-pdbzkoijyr-0-ppo4rdunfyhl-kube-minion-agodggladxws Ready

kubectl create -f /etc/kubernetes/examples/pod-nginx-with-label.yaml
kubectl get pods
NAME READY STATUS RESTARTS AGE
nginx 1/1 Running 0 5m

kubectl exec -ti nginx -- bash
error: Unable to upgrade connection: {
  "kind": "Status",
  "apiVersion": "v1",
  "metadata": {},
  "status": "Failure",
  "message": "dial tcp: lookup k8-pdbzkoijyr-0-ppo4rdunfyhl-kube-minion-agodggladxws: no such host",
  "code": 500
}

[Ton Ngo (ton-i)]

The script adjusts the host name to match the nova instance name. One solution would be to adjust the DNS entry accordingly.

[Ton Ngo (ton-i)]

Remember to update the Troubleshooting guide to use "kubectl exec" instead of "docker exec" so that users don't need to log into the node.

[Mark Goddard (mgoddard)]

Running into this myself. It's possible to get nova to register instance hostnames with neutron's internal dns: https://docs.openstack.org/mitaka/networking-guide/config-dns-int.html.