CoreDNS in LKE

Linode Kubernetes Engine (LKE) provides out of the box intra-cluster domain name resolution via CoreDNS, the DNS server. Every new cluster is provided with a minimal, default CoreDNS configuration, which can be customized to suit your workload's needs.

Before you begin

This guide assumes you have a working Linode Kubernetes Engine (LKE) cluster running on Linode and you are familiar with Corefile, the CoreDNS configuration file.

1. Install the Kubernetes CLI (kubectl) on the local computer.

2. Follow the instructions in Deploying and Managing a Cluster with Linode Kubernetes Engine Tutorial to connect to an LKE cluster.

   📘

   Ensure that the KUBECONFIG context is persistent

3. Ensure that Kubernetes CLI is using the right cluster context. Run the get-contexts subcommand to check:

   kubectl config get-contexts
   

Default CoreDNS configuration

You can view your cluster's default CoreDNS configuration by using the following command:

kubectl get configmap -n kube-system coredns-base -o yaml

The output will resemble the following:

apiVersion: v1
data:
  Corefile: |
    .:53 {
        errors
        health {
           lameduck 5s
        }
        ready
        kubernetes cluster.local in-addr.arpa ip6.arpa {
           pods insecure
           fallthrough in-addr.arpa ip6.arpa
           ttl 30
        }
        prometheus :9153
        forward . /etc/resolv.conf {
           max_concurrent 1000
        }
        cache 30
        loop
        reload
        loadbalance
        import custom/*.include
    }
    import custom/*.server
kind: ConfigMap
metadata:
  name: coredns-base
  namespace: kube-system
  [...]

The default CoreDNS configuration is located under the Corefile field in the above ConfigMap.

🚧

Do not modify the kube-system/coredns-base ConfigMap that comes with your LKE cluster. It may be restored to its original state at any time and without notice.

Custom CoreDNS configuration

The default CoreDNS configuration leverages the CoreDNS import plugin to enable customization. Configuration extensions are added through fields in the kube-system/coredns-custom ConfigMap:

    apiVersion: v1
    kind: ConfigMap
    metadata:
      name: coredns-custom
      namespace: kube-system
    data:
      sample.include: |
        # Added to the .:53 default Server Block.
      sample_a.server: |
        # Additional Server Block.
      sample_b.server: |
        # Another Server Block.

• Fields suffixed with .include are added to the default Server Block.
• Fields suffixed with .server are added as new Server Blocks.

Create a custom configuration

1. Create a manifest for a ConfigMap named coredns-custom in the kube-system namespace, with the desired configuration. For the purpose of this guide, an example custom configuration is used. Save it as the coredns-custom.yaml file.

   YAML

   apiVersion: v1
   kind: ConfigMap
   metadata:
     name: coredns-custom
     namespace: kube-system
   data:
     # Log all incoming DNS queries.
     log.include: |
       log
     # Private DNS resolution example. Handles FQDN resolutions for *.mydomain.com
     # Replace <dns-private-ip> with the target IP address.
     mydomain.server: |
       mydomain.com.:53 {
         forward . <dns-private-ip>
       }
   

2. Apply the above ConfigMap manifest:

   kubectl apply -f coredns-custom.yaml
   

   📘

   CoreDNS will attempt to reload the configuration within 45 seconds after the last modification.

3. Ensure the custom configuration has been loaded:

   kubectl logs -n kube-system -l k8s-app=kube-dns
   

   For the custom configuration shown above, the output will resemble the following, after the reload is complete:

   Output

   [INFO] Reloading
   [INFO] plugin/health: Going into lameduck mode for 5s
   [INFO] 127.0.0.1:60399 - 40866 "HINFO IN 349145763287755047.2816822520842364744. udp 56 false 512" NXDOMAIN qr,rd,ra 131 0.000980597s
   [INFO] plugin/reload: Running configuration SHA512 = 868c96ccca274c442fefc8db8e98b1f4a5cd05c655db1d990803d4019e5d28af101b24a78f85bae7ab3a3f8894f2791fda9d2b4d9c6ae1aa942080e1a88ce3e6
   [INFO] Reloading complete
   

   The custom configuration is now in effect.

📘

The kube-system/coredns-custom ConfigMap is persistent and will not be affected by LKE maintenance operations.

Restore the defaults

1. To restore the default CoreDNS configuration, simply delete the coredns-custom ConfigMap:

   kubectl delete -n kube-system coredns-custom
   

2. Check the logs to make sure the reload was successful:

   kubectl logs -n kube-system -l k8s-app=kube-dns
   

   The output looks similar to the one emitted after applying the custom configuration.

   Output

   [INFO] Reloading
   [INFO] plugin/health: Going into lameduck mode for 5s
   [WARNING] No files matching import glob pattern: custom/*.include
   [WARNING] No files matching import glob pattern: custom/*.server
   [INFO] plugin/reload: Running configuration SHA512 = 591cf328cccc12bc490481273e738df59329c62c0b729d94e8b61db9961c2fa5f046dd37f1cf888b953814040d180f52594972691cd6ff41be96639138a43908
   [INFO] Reloading complete
   

   The emitted warning messages are now to be expected, and should not be a concern.