Upgrade
Keep kube-vip on a current release. You can upgrade an existing cluster in place;
you do not need to rebuild control-plane nodes or re-run kubeadm init.
What changes on upgrade
Upgrades replace the kube-vip container image (and optionally regenerate the manifest when flags or defaults have changed between versions). Leader election handles the hand-off: when a kube-vip Pod restarts, another instance can take the VIP, then the upgraded instance rejoins election as usual.
Choose your install path
| How kube-vip was installed | Upgrade guide |
|---|---|
Static Pod under /etc/kubernetes/manifests (typical kubeadm) | Upgrade using manifest |
| DaemonSet (K3s and similar) | Follow the DaemonSet steps on the manifest upgrade page |
Before you start
- Note the VIP, interface name, and feature flags you use today (ARP/BGP, control plane, services, leader election). You will need the same values if you regenerate the manifest.
- Pick a target version from the kube-vip releases page.
- For production clusters, upgrade one control-plane node at a time and confirm the API server remains reachable on the VIP after each node.