[Bug]: Nodes unable to start/connect after fresh creation about terraform-hcloud-kube-hetzner HOT 4 CLOSED

I had to delete the state, in my case an S3 bucket, because there was always something wrong when creating the cluster. I was not able to import the resources and destroying failed due to expected resources that did just not exist, such as a remote kubeconfig for example

from terraform-hcloud-kube-hetzner.

I resolved the issue, it wasn't a really terraform issue. I added another snapshot image with accidently the same labels as used by k3s. So deleting this snapshot and deleting the cloudinit from tf state via terraform state rm module.k8s.data.cloudinit_config.autoscaler_config[0] and then running terraform apply resolved the issue

from terraform-hcloud-kube-hetzner.

As pointed out by Brad Davidson in k3s-io/k3s#9343, the issue was related to a faulty k3s configuration. All nodes were pointing towards each other with the ca certs without any of the nodes actually exposing them. This is most likely related to a faulty state, as this indicates the script expected to join an existing cluster instead of creating one from scratch. Deleting the state instead of just destroying it solved the immediate issue.

from terraform-hcloud-kube-hetzner.

@yannicschroeer did you just deleted the .terraform folder or did you complete destroy all resources and then deleting also the .terraform folder?
I'm having the same issue and create a new terraform state and looking how I can "re-import" my cluster resources...

from terraform-hcloud-kube-hetzner.