x509: certificate is valid for X, not Y (Kubernetes) — how to fix
Unable to connect to the server: x509: certificate is valid for 10.96.0.1, 172.31.5.10, not 54.210.33.12
tls: failed to verify certificate: x509: certificate is valid for kubernetes, kubernetes.default, not k8s.example.com
Quick fix: Either connect using a name/IP that's in the cert (check with `openssl s_client -connect HOST:6443
What it means
TLS verification checks two things: that the certificate is signed by a trusted CA, and that the address you connected to appears in the certificate's SAN list. The CA part passed here; the name check failed. The message lists the names the cert is valid for, then the one you used (not ...).
Typical situations:
- The cluster was built with private IPs and you now connect through a public IP or NAT (cloud VM, home lab port-forward).
- A load balancer or DNS name (
k8s.example.com) was added in front of the control plane later. - The node's IP changed after the cert was issued.
See the SANs of the certificate actually served:
openssl s_client -connect 54.210.33.12:6443 </dev/null 2>/dev/null \
| openssl x509 -noout -ext subjectAltName
And check which address your kubeconfig uses:
kubectl config view --minify -o jsonpath='{.clusters[0].cluster.server}'
Fix 1: use an address that is in the certificate
If the cert includes a DNS name or IP you can reach (e.g. through a VPN), point kubeconfig at it:
kubectl config set-cluster mycluster --server=https://172.31.5.10:6443
If only the name is wrong but the IP works, you can keep the IP and override the name used for verification:
kubectl config set-cluster mycluster --tls-server-name=kubernetes
tls-server-name (kubectl 1.19+) keeps full verification, unlike skipping TLS.
Fix 2: kubeadm, add SANs and regenerate the API server cert
On a control-plane node, add the names to the kubeadm config:
# kubeadm-config.yaml (ClusterConfiguration)
apiServer:
certSANs:
- "54.210.33.12"
- "k8s.example.com"
Then regenerate (back up first):
sudo mkdir -p /root/pki-backup && sudo cp -a /etc/kubernetes/pki /root/pki-backup/
sudo rm /etc/kubernetes/pki/apiserver.{crt,key}
sudo kubeadm init phase certs apiserver --config kubeadm-config.yaml
The kube-apiserver static pod picks up the new cert after a restart; the simplest way is to move its manifest out and back:
sudo mv /etc/kubernetes/manifests/kube-apiserver.yaml /tmp/ && sleep 20
sudo mv /tmp/kube-apiserver.yaml /etc/kubernetes/manifests/
For a new cluster, pass the SANs up front: kubeadm init --apiserver-cert-extra-sans=54.210.33.12,k8s.example.com. Repeat the cert step on every control-plane node.
Fix 3: k3s / RKE2
Add the SAN to the config and restart; k3s regenerates the serving cert automatically:
# /etc/rancher/k3s/config.yaml
tls-san:
- 54.210.33.12
- k8s.example.com
sudo systemctl restart k3s
(RKE2 uses /etc/rancher/rke2/config.yaml and rke2-server.)
Fix 4: managed clusters and minikube
On EKS/GKE/AKS you can't change the API cert; always use the endpoint from aws eks update-kubeconfig, gcloud container clusters get-credentials or az aks get-credentials. For minikube use minikube start --apiserver-ips=... --apiserver-names=....
Same error outside kubectl
x509: certificate is valid for ..., not ... from Docker, Go programs or curl means the same thing for any server: the hostname you use must be in the cert. Fix the cert (e.g. reissue with Let's Encrypt including all names) or connect with the right hostname.
Avoid insecure-skip-tls-verify: true in kubeconfig: it disables all certificate checks, not just the name.
Last updated 2026-11-11.