HKS kube-vip

HKS can use the product-managed Kube VIP load balancer to provide a virtual IP address for a Kubernetes cluster. Kube VIP is available for Kubernetes Cluster layouts using VMware, vCloud Director, Xen, manual, KVM, and standard provision types.

Prerequisites and Configuration

Select Kube VIP as the cluster load balancer and provide either:

  • A VIP Pool from which HPE Morpheus Software acquires an address. The address must be accessible from every cluster node.

  • A VIP Address that is unused and routable on the node network.

HPE Morpheus Software owns kube-vip configuration. During node configuration, it selects the appropriate master or worker script and writes the generated configuration to the node. Master nodes run kube-vip as a static Pod with leader election. Do not edit the generated /etc/kubernetes/manifests/kube-vip.yaml file; later HPE Morpheus Software operations can replace local changes.

Verification and Troubleshooting

  1. Confirm the cluster load balancer shows Kube VIP and the expected virtual IP.

  2. From a system on the node network, verify that the virtual IP is reachable.

  3. On each control-plane node, inspect the kube-system namespace and confirm its kube-vip-<node> Pod is running.

  4. Verify the generated static Pod manifest references the expected image and virtual IP without modifying it.

If the virtual IP is unavailable, verify that it is unused, belongs to a network reachable by all nodes, and is not blocked by network policy or switching configuration. Review the cluster provisioning process output for the Kube VIP node-configuration step. On a multi-control-plane cluster, loss of the leader should cause another kube-vip Pod to acquire leadership; verify the virtual IP remains reachable before returning the cluster to service.