CoreDNS

CoreDNS is a DNS server that chains plugins and provides DNS Services

TL;DR;

$ helm repo add coredns https://coredns.github.io/helm
$ helm --namespace=kube-system install coredns coredns/coredns

Introduction

This chart bootstraps a CoreDNS deployment on a Kubernetes cluster using the Helm package manager. This chart will provide DNS Services and can be deployed in multiple configuration to support various scenarios listed below:

• CoreDNS as a cluster dns service and a drop-in replacement for Kube/SkyDNS. This is the default mode and CoreDNS is deployed as cluster-service in kube-system namespace. This mode is chosen by setting isClusterService to true.
• CoreDNS as an external dns service. In this mode CoreDNS is deployed as any kubernetes app in user specified namespace. The CoreDNS service can be exposed outside the cluster by using using either the NodePort or LoadBalancer type of service. This mode is chosen by setting isClusterService to false.
• CoreDNS as an external dns provider for kubernetes federation. This is a sub case of 'external dns service' which uses etcd plugin for CoreDNS backend. This deployment mode as a dependency on etcd-operator chart, which needs to be pre-installed.

Prerequisites

• Kubernetes 1.10 or later

Installing the Chart

The chart can be installed as follows:

$ helm repo add coredns https://coredns.github.io/helm
$ helm --namespace=kube-system install coredns coredns/coredns

The command deploys CoreDNS on the Kubernetes cluster in the default configuration. The configuration section lists various ways to override default configuration during deployment.

Tip: List all releases using helm list --all-namespaces

Uninstalling the Chart

To uninstall/delete the coredns deployment:

$ helm uninstall coredns

The command removes all the Kubernetes components associated with the chart and deletes the release.

Configuration

See values.yaml for configuration notes. Specify each parameter using the --set key=value[,key=value] argument to helm install. For example,

$ helm install coredns \
  coredns/coredns \
  --set rbac.create=false

The above command disables automatic creation of RBAC rules.

Alternatively, a YAML file that specifies the values for the above parameters can be provided while installing the chart. For example,

$ helm install coredns coredns/coredns -f values.yaml

Tip: You can use the default values.yaml

Caveats

The chart will automatically determine which protocols to listen on based on the protocols you define in your zones. This means that you could potentially use both "TCP" and "UDP" on a single port. Some cloud environments like "GCE" or "Azure container service" cannot create external loadbalancers with both "TCP" and "UDP" protocols. So When deploying CoreDNS with serviceType="LoadBalancer" on such cloud environments, make sure you do not attempt to use both protocols at the same time.

Autoscaling

By setting autoscaler.enabled = true a cluster-proportional-autoscaler will be deployed. This will default to a coredns replica for every 256 cores, or 16 nodes in the cluster. These can be changed with autoscaler.coresPerReplica and autoscaler.nodesPerReplica. When cluster is using large nodes (with more cores), coresPerReplica should dominate. If using small nodes, nodesPerReplica should dominate.

This also creates a ServiceAccount, ClusterRole, and ClusterRoleBinding for the autoscaler deployment.

replicaCount is ignored if this is enabled.

By setting hpa.enabled = true a Horizontal Pod Autoscaler is enabled for Coredns deployment. This can scale number of replicas based on meitrics like CpuUtilization, MemoryUtilization or Custom ones.

Adopting existing CoreDNS resources

If you do not want to delete the existing CoreDNS resources in your cluster, you can adopt the resources into a release as of Helm 3.2.0.

You will also need to annotate and label your existing resources to allow Helm to assume control of them. See: https://github.com/helm/helm/pull/7649

annotations:
  meta.helm.sh/release-name: your-release-name
  meta.helm.sh/release-namespace: your-release-namespace
label:
  app.kubernetes.io/managed-by: Helm

Once you have annotated and labeled all the resources this chart specifies, you may need to locally template the chart and compare against existing manifest to ensure there are no changes/diffs.s If you have been careful this should not diff and leave all the resources unmodified and now under management of helm.

Some values to investigate to help adopt your existing manifests to the Helm release are:

• k8sAppLabelOverride
• service.name
• customLabels

In some cases, you will need to orphan delete your existing deployment since selector labels are immutable.

kubectl delete deployment coredns --cascade=orphan

This will delete the deployment and leave the replicaset to ensure no downtime in the cluster. You will need to manually delete the replicaset AFTER Helm has released a new deployment.

Here is an example script to modify the annotations and labels of existing resources:

WARNING: Substitute YOUR_HELM_RELEASE_NAME_HERE with the name of your helm release.

#!/usr/bin/env bash

set -euo pipefail

for kind in config service serviceAccount; do
    echo "setting annotations and labels on $kind/coredns"
    kubectl -n kube-system annotate --overwrite $kind coredns meta.helm.sh/release-name=YOUR_HELM_RELEASE_NAME_HERE
    kubectl -n kube-system annotate --overwrite $kind coredns meta.helm.sh/release-namespace=kube-system
    kubectl -n kube-system label --overwrite $kind coredns app.kubernetes.io/managed-by=Helm
done

NOTE: Sometimes, previous deployments of kube-dns that have been migrated to CoreDNS still use kube-dns for the service name as well.

echo "setting annotations and labels on service/kube-dns"
kubectl -n kube-system annotate --overwrite service kube-dns meta.helm.sh/release-name=YOUR_HELM_RELEASE_NAME_HERE
kubectl -n kube-system annotate --overwrite service kube-dns meta.helm.sh/release-namespace=kube-system
kubectl -n kube-system label --overwrite service kube-dns app.kubernetes.io/managed-by=Helm