Deploying NodeLocal DNSCache to Improve DNS Resolution in Kubernetes Clusters

During the operation of a Kubernetes (K8s) cluster, some Pods may experience request timeouts when resolving Service names. This occurs because DNS queries from Pods are handled by CoreDNS, which can become a bottleneck in large‑scale clusters.

The Kubernetes community introduced the NodeLocal DNSCache solution to enhance DNS resolution performance by deploying a DNS cache on each node using a DaemonSet, thereby offloading CoreDNS and reducing network latency for Pods.

1. Solution Overview

NodeLocal DNSCache runs a DNS caching service on every node, avoiding cross‑node network hops and alleviating CoreDNS pressure. After enabling it, DNS queries follow a flow where cached responses are served locally, and only uncached queries are forwarded to CoreDNS.

2. Deploy NodeLocal DNSCache

1. Create a resource manifest named nodelocaldns.yaml (example can be found at GitHub ).

2. Modify the manifest variables to match your environment:

kubedns=`kubectl get svc kube-dns -n kube-system -o jsonpath={.spec.clusterIP}`

domain=<cluster-domain>
localdns=<node-local-address>

# Example values

domain=cluster.local
localdns=169.254.20.10

3. Adjust settings based on the kube‑proxy mode:

If kube‑proxy runs in IPTABLES mode, execute: sed -i "s/__PILLAR__LOCAL__DNS__/$localdns/g; s/__PILLAR__DNS__DOMAIN__/$domain/g; s/__PILLAR__DNS__SERVER__/$kubedns/g" nodelocaldns.yaml

If kube‑proxy runs in IPVS mode, execute: sed -i "s/__PILLAR__LOCAL__DNS__/$localdns/g; s/__PILLAR__DNS__DOMAIN__/$domain/g; s/,__PILLAR__DNS__SERVER__//g; s/__PILLAR__CLUSTER__DNS__/$kubedns/g" nodelocaldns.yaml

4. Deploy the manifest:

kubectl create -f nodelocaldns.yaml

After deployment, the node‑local‑dns Pods run in the kube-system namespace on each node.

$ kubectl get pod -n kube-system -l k8s-app=node-local-dns
NAME                     READY   STATUS    RESTARTS   AGE
node-local-dns-clrsb     1/1     Running   0          6m35s
node-local-dns-msx22    1/1     Running   0          6m35s
node-local-dns-wpphv     1/1     Running   0          6m35s

If using IPVS mode, update each node’s kubelet --cluster-dns flag to point to the local DNS address, then reload and restart the kubelet:

sed -i 's/$kubedns/$localdns/g' /var/lib/kubelet/config.yaml
systemctl daemon-reload && systemctl restart kubelet

Verify the new DNS configuration on a node:

$ cat /etc/resolv.conf
nameserver 169.254.20.10
search default.svc.cluster.local svc.cluster.local cluster.local
options ndots:5

At this point, the deployment is complete and you can experience the performance benefits of NodeLocal DNSCache.