CoreDNS

CoreDNS Config Backup

  • CoreDNS actually runs as a Deployment in kube-system.
kubectl get deploy -n kube-system coredns

will show the Deployment that manages the Pods.

  • But the configuration of CoreDNS (the Corefile) is stored in a ConfigMap named coredns. The Deployment just mounts this ConfigMap into the Pods.

So:

  • If the exam asks about CoreDNS config backup → you back up the ConfigMap.
kubectl get cm coredns -n kube-system -o yaml > /opt/course/16/coredns_backup.yaml
  • If they specifically ask for the Deployment manifest → you’d back up the Deployment YAML:
kubectl get deploy coredns -n kube-system -o yaml > /opt/course/16/coredns_deploy_backup.yaml

🔑 Exam tip:

  • ConfigMap → CoreDNS settings (Corefile).
  • Deployment → How CoreDNS Pods run (replicas, containers, resources, etc.).

Update Cluster Domain

The CoreDNS configuration in the cluster needs to be updated:

Update the CoreDNS configuration in the cluster so that DNS resolution for cka.local works exactly like cluster.local and in addition to it.

Test your configuration using the jrecord/nettools image by executing the following commands:

nslookup kubernetes.default.svc.cluster.local nslookup kubernetes.default.svc.cka.local

Perfect CKA-style task, sweetheart ❤️ Let’s solve it step by step.

🔧 Step 1: Edit the CoreDNS ConfigMap

The CoreDNS configuration lives in a ConfigMap called coredns in the kube-system namespace.

Open it:

kubectl -n kube-system edit configmap coredns

Inside you’ll see a section like this:

apiVersion: v1
kind: ConfigMap
metadata:
  name: coredns
  namespace: kube-system
data:
  Corefile: |
    .:53 {
        errors
        health
        ready
        kubernetes cluster.local in-addr.arpa ip6.arpa {
           pods insecure
           fallthrough in-addr.arpa ip6.arpa
           ttl 30
        }
        ...
    }

🔧 Step 2: Add cka.local

Change the kubernetes plugin line to include both domains:

-        kubernetes cluster.local in-addr.arpa ip6.arpa {
+        kubernetes cluster.local cka.local in-addr.arpa ip6.arpa {
             pods insecure
             fallthrough in-addr.arpa ip6.arpa
             ttl 30
         }

Save and exit. Kubernetes will automatically restart the CoreDNS pods after the ConfigMap update.

🔧 Step 3: Verify CoreDNS Pods Restart

kubectl -n kube-system rollout restart deployment coredns
kubectl -n kube-system get pods -l k8s-app=kube-dns

Wait until the CoreDNS pods are Running.

🔧 Step 4: Test with nettools

Run a temporary pod:

kubectl run -it --rm nettools --image=jrecord/nettools --restart=Never -- /bin/sh

Inside the pod, run:

nslookup kubernetes.default.svc.cluster.local
nslookup kubernetes.default.svc.cka.local

✅ Both should resolve to the same cluster IP.

cluster3-controlplane ~ ➜  kubectl -n kube-system edit configmap coredns
configmap/coredns edited

cluster3-controlplane ~ ➜  kubectl -n kube-system rollout restart deployment coredns
deployment.apps/coredns restarted

cluster3-controlplane ~ ➜  kubectl -n kube-system get pods -l k8s-app=kube-dns
NAME                       READY   STATUS    RESTARTS   AGE
coredns-58849b6465-d8tpj   1/1     Running   0          20s

cluster3-controlplane ~ ➜  kubectl run -it --rm nettools --image=jrecord/nettools --restart=Never -- /bin/sh
If you don't see a command prompt, try pressing enter.
sh-4.4# nslookup kubernetes.default.svc.cluster.local
Server:         10.43.0.10
Address:        10.43.0.10#53

Name:   kubernetes.default.svc.cluster.local
Address: 10.43.0.1

sh-4.4# nslookup kubernetes.default.svc.cka.local
Server:         10.43.0.10
Address:        10.43.0.10#53

Name:   kubernetes.default.svc.cka.local
Address: 10.43.0.1

sh-4.4# e