core-dns failing in civo cluster

[ratnadeep007]

Issue Unable to create vcluster in Civo kubernetes cluster which internally uses k3s. Able to connect to api-server but unable to deploy anything as coredns is failing for vcluster

Steps

1. Created namespace; kubectl create namespace vcluster-1
2. Used vcluster cli: vcluster create vcluster-1 -n vcluster-1
3. Connect to cluster: vcluster connect vcluster-1 -n vcluster-1
4. Tried deploying nginx, it stays in pending state inside vcluster but running state on host.
5. coredns container is in Pending state for vcluster and ContainerCreating state for host

Logs Events from describe command of core-dns

• Unable to attach or mount volumes: unmounted volumes=[config-volume], unattached volumes=[config-volume coredns-token-mbwt2]: timed out waiting for the condition
• MountVolume.SetUp failed for volume "config-volume" : configmap references non-existent config key: NodeHosts

Self fix or debug Tried changing default storageclass in cluster from civo-volume from provisioner csi.civo.com to local-path from provisioner rancher.io/local-path but same result.

Cluster Type: Managed Cluster based on k3s by Civo.com

[FabianKramm]

@ratnadeep007 thanks for creating this issue! Thats odd, this usually only happens if you try to run vcluster in a cluster without any nodes. What happens if you deploy vcluster with a values.yaml that looks like this:

storage:
  persistence: false
rbac:
  clusterRole:
    create: true
syncer:
  extraArgs:
  - --sync-all-nodes
  - --fake-nodes=false
  - --fake-kubelets=false

and then create the vcluster with vcluster create vcluster-1 -n vcluster-1 -f values.yaml, does that work?

[ratnadeep007]

@FabianKramm Thanks, it worked. Closing issue.