The connection to the server localhost:8080 was refused
kubectl didn’t find a usable kubeconfig, so it tried its built-in default of localhost:8080, where nothing is listening.
Diagnose
Section titled “Diagnose”echo "$KUBECONFIG"ls -l ~/.kube/configkubectl config current-contextUsual causes
Section titled “Usual causes”| Cause | Fix |
|---|---|
No ~/.kube/config for this user |
Copy or generate the kubeconfig (for kubeadm, from /etc/kubernetes/admin.conf; for managed clusters, the provider’s CLI) |
Running as a different user, often with sudo |
sudo uses root’s home directory; run without it, or set KUBECONFIG |
KUBECONFIG points at a missing file |
Fix the path |
| Current context points at nothing | kubectl config use-context <name> |
On a control plane node, right after kubeadm init |
Follow the printed instructions to copy admin.conf to ~/.kube/config |
admin.conf grants full cluster-admin access. Don’t copy it to laptops or CI systems; use per-person credentials from your identity provider instead.