DNS server validation fails

Installation fails at the validating DNS server stage.

Symptoms

Installation fails with the following message:

fatal: [x.x.x.x] => A loopback IP is used in your DNS server configuration. For more details, see https://ibm.biz/dns-fails.

Causes

The cluster node that is specified in the error message uses a loopback IP(127.0.0.1 or 127.0.1.1) as the DNS server. Or, the cluster node that is specified in the error message does not have a /etc/resolv.conf file.

When a loopback IP is used as the DNS server, you might experience issues when a service tries to resolve external domain from inside a container.

Resolving the problem

Add the following setting to the cluster/config.yaml configuration file.
```
 loopback_dns: true
```
Retry the installation.

After you successfully install your cluster, update the kube-dns configuration.

 kubectl edit cm kube-dns --namespace=kube-system

     apiVersion: v1
 data:
   Corefile: |
     .:53 {
         errors
         health
         kubernetes cluster.local in-addr.arpa ip6.arpa {
             pods insecure
             upstream
             fallthrough in-addr.arpa ip6.arpa
         }
         prometheus :9153
  -        proxy . /etc/resolv.conf
 +        proxy . 8.8.8.8
         cache 30
         reload
     }
 kind: ConfigMap
 metadata:
   creationTimestamp: 2018-08-22T01:54:16Z
   labels:
     addonmanager.kubernetes.io/mode: EnsureExists
     app: kube-dns
     chart: kube-dns-99.99.99
     heritage: Tiller
     release: kube-dns
   name: kube-dns
   namespace: kube-system
   resourceVersion: "6382"
   selfLink: /api/v1/namespaces/kube-system/configmaps/kube-dns
   uid: 46c0e47f-a5ae-11e8-9bba-fa163eec7c9e

Restart the kube-dns pod to load the new configuration.

 [root@kvm-014377 ~]# kubectl get po -n kube-system |grep kube-dns
 kube-dns-ldtqc                                                1/1       Running     0          30m
 [root@kvm-014377 ~]# kubectl delete po kube-dns-ldtqc -n kube-system
 pod "kube-dns-ldtqc" deleted