Why did I specify clusterIP for CoreDNS?

[rabbagliettiandrea]

Hi everyone,

I was reviewing the foundation configurations of one of our production clusters, and I noticed something in the values.yaml for CoreDNS. Specifically, I saw that I had manually set the clusterIP like this:

replicaCount: 2  
service:  
  clusterIP: "10.100.0.10"  

The issue I'm running into now is that I'm unable to upgrade CoreDNS because it says the IP is already bound.

The problem is... I honestly can't remember why I set the clusterIP in the first place.

Does anyone have any ideas? Could this have something to do with not using KubeDNS? I'm a bit confused.

Here's my kubectl get all -n kube-system:

[vinaycharlie01]

The problem is... I honestly can't remember why I set the clusterIP in the first place.

Currently, I'm working with CoreDNS in a K3s setup, and in my case, the default cluster IP is 10.43.0.1. So I’ve set the DNS cluster IP to 10.43.0.10.

You can verify your default cluster IP using:

kubectl get svc kubernetes -n default -o jsonpath='{.spec.clusterIP}'

It might be the same in your case too.

The issue I'm running into now is that I'm unable to upgrade CoreDNS because it says the IP is already bound.

If you're running into an issue upgrading due to the IP being already bound, you can follow these steps:

# 1. Delete the existing CoreDNS service
kubectl delete svc coredns -n kube-system

# 2. Upgrade CoreDNS using Helm with the desired clusterIP
helm upgrade coredns coredns/coredns \
  --namespace kube-system \
  --set service.clusterIP=<<IP>>

[vinaycharlie01]

@rabbagliettiandrea

[rabbagliettiandrea]

@vinaycharlie01 Thank you so much! I’ve arrived at more or less the same solution, but without using service.clusterIP from the Helm chart, since I don't see that key existing. Based on what I found here: https://artifacthub.io/packages/helm/coredns/coredns?modal=values — maybe I’m wrong?

[gricuk]

@rabbagliettiandrea 10.100.0.10 is default IP assigned as clusterDNS in EKS for example.