Unable to remove kube-dns from Kubernetes Cluster v1.9.1 about deployment

[campbelldgunn]

@johnbelamaric Any ideas would be greatly appreciated as I was just following the instructions. It is a v1.9.1 K8s cluster, so RBAC is enable. thx.

[johnbelamaric]

It is probably the add-on manager. How did you build the cluster? Is it a managed cluster like GKE? On Jan 30, 2018, at 4:53 AM, Campbell Gunn <[email protected]<mailto:[email protected]>> wrote: @johnbelamaric<https://github.com/johnbelamaric> Any ideas would be greatly appreciated as I was just following the instructions. It is a v1.9.1 K8s cluster, so RBAC is enable. thx. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#59 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AJB4s5qvjTkl4ahf-39Dr-nIZKH5xXEHks5tPuatgaJpZM4Rx6ks>.

[campbelldgunn]

The cluster was built using acs-engine in MS Azure.

[chrisohaver]

It does sound like an addon-manager. I have no direct experience with Azure deployments of kubernetes. But inferring from some other discussions, you might try the following ...

Look in /etc/kubernetes/addons/ on the master for a kube-dns yaml definition. Try deleting that yaml file (or moving it out of the directory), then restarting the addon manager (by deleting the addon-manager pod in kube-system namespace). Then try deleting the kube-dns deployment.

[chrisohaver]

... Also be aware that the addon manager may be reconciling the kube-dns service too (i.e. restoring it to select the kube-dns pods instead of the coredns pods). If so, you'll need to re-apply the yaml produced by the deployment.sh script.

[campbelldgunn]

All right I will have a look at that. thanks.

[campbelldgunn]

@chrisohaver Nope, did exactly the steps above and the Kube-DNS deployment comes back and so do the pods.