VMware vSphere Provider

Prerequisites

Create a Cluster

Delete a Cluster

Troubleshooting

Reference

The Infrastructure Provider for Virtualization Platform creates Alauda OS clusters on VMware vSphere: it clones every node from a VM template through vCenter. When you plan a cluster, note what is specific to vSphere:

  • Every node has a hostname and a static IPv4 address that you list in install-config.yaml. There is no DHCP, and each node group has its own pool of these slots.
  • The control plane endpoint is your external load balancer, the default, or a Self-built VIP that the provider runs on the control plane nodes. The choice is fixed when the cluster is created.
  • /var/cpaas and raw data disks are persistent: when a VM is replaced, they are reattached to its replacement. The containerd, kubelet, and etcd disks are recreated with the VM.
  • A released persistent disk is deleted after releaseDelayHours. Delete a cluster with --reclaim-disks to delete its persistent disks with it.
  • The platform upgrade also upgrades Kubernetes and Alauda OS on these clusters; see Who Upgrades Kubernetes.

How It Works

VMware vSphere Cluster API resource dependencies

The vSphere cloud provider is delivered to the cluster automatically; the nodes stay uninitialized until it runs.

Documentation

Read the pages in this order:

  1. Prerequisites, and Install the Provider when the global cluster does not run it yet
  2. Create a Cluster with ACP CLI (ac)
  3. Troubleshooting, when a creation stage does not advance
  4. Delete a Cluster

Related: Parameters, Manage Nodes, Upgrade Overview, Provider APIs.