CoreDNS Component Manual Dilatation Guide

Compatibility of CoreDNS and K8S version

⚠️ Note: Upgrading CoreDNS will lead to a 5-minute network interruption across the entire cluster. Please proceed with the upgrade only after a comprehensive evaluation.

Upgrade process

1. To check the current CoreDNS image version in the cluster, use the following command:
   kubectl -n kube-system get deployment coredns -o yaml | grep "image:"
   Example output:
   image: registry.baidubce.com/cce-plugin-pro/coredns:1.7.1
2. Modify potential plugin name changes.
   If the CoreDNS image version is very old (below 1.6.2, where the forward plugin was previously named proxy), use the following command:
   kubectl -n kube-system get cm coredns -o yaml
   If the plugin name shown in the diagram is not forward but proxy:
   
   modify it to forward
   using the following commandkubectl -n kube-system get cm coredns -o yaml
3. add endpointslices permissions
   Use the following command:
   kubectl -n kube-system get clusterrole system:coredns -o yaml
   If the result shows missing endpointslices permissions, proceed via:
   kubectl -n kube-system edit clusterrole system:coredns
   Add permissions. The modified clusterrole content is as follows:

1apiVersion: rbac.authorization.k8s.io/v1
2kind: ClusterRole
3metadata:
4  labels:
5    kubernetes.io/bootstrapping: rbac-defaults
6  name: system:coredns
7rules:
8  - apiGroups:
9    - ""
10    resources:
11    - endpoints
12    - services
13    - pods
14    - namespaces
15    verbs:
16    - list
17    - watch
18  - apiGroups:
19    - discovery.k8s.io
20    resources:
21    - endpointslices
22    verbs:
23    - list
24    - watch

4. Modify image version Upgrade the CoreDNS image version to the desired one (e.g., 1.9.4-baidu) with the following command:
   kubectl -n kube-system edit deployment coredns The modified example is as follows:
   
5. Check upgrade result Use the command:
   kubectl -n kube-system get pod -o wide | grep coredns
   Check whether all pods have successfully started Use the command:
   kubectl -n kube-system logs <coredns-pod-name> | less -N
   Check whether the CoreDNS pod has error logs
   If abnormalities exist in the above process, submit a ticket for handling; if all are normal, the upgrade is completed.