Prerequisites and limitations

Prerequisites and limitations for configuring a vSphere Cloud Provider in your IBM® Cloud Private cluster.

Prerequisites

Make sure that the nodes in your cluster meet these requirements:

All IBM® Cloud Private cluster nodes must be under one vSphere VM folder that is under the root folder. The nodes must not be in the root folder.
All IBM Cloud Private master nodes must be able to access vCenter.
The node host name must be same as the VM name.
Node host names must comply with the regex, [a-z](([-0-9a-z]+)?[0-9a-z])?(\.[a-z0-9](([-0-9a-z]+)?[0-9a-z])?)*, and must also comply with the following restrictions:
- They must not begin with numbers.
- They must not use capital letters.
- They must not have any special characters except . and -.
- They must contain at least three characters but no more than 63 characters.
The disk UUID on the node VMs must be enabled: the disk.EnableUUID value must be set to True.
The user that is specified in the vSphere cloud configuration must have privileges to interact with vCenter.

Table 1. vSphere Cloud Provider user
Roles	Privileges	Entities	Propagate to child
manage-k8s-node-vms	Resource.AssignVMToPool System.Anonymous System.Read System.View VirtualMachine.Config.AddExistingDisk VirtualMachine.Config.AddNewDisk VirtualMachine.Config.AddRemoveDevice VirtualMachine.Config.RemoveDisk VirtualMachine.Inventory.Create VirtualMachine.Inventory.Delete	Cluster Hosts VM Folder	Yes
manage-k8s-volumes	Datastore.AllocateSpace Datastore.FileManagement System.Anonymous System.Read System.View	Datastore	No
k8s-system-read-and-spbm-profile-view	StorageProfile.View System.Anonymous System.Read System.View	vCenter	No
ReadOnly	System.Anonymous System.Read System.View	Datacenter Datastore Cluster Datastore Storage Folder	No

vSphere Cloud Provider supports only the ReadWriteOnce access mode for persistent volumes.
If a node is down, all the pods along with the mounted volumes move to another node. However, the old node does not come back up because the volumes are not automatically unmounted. You must manually unmount the volumes from the old node for it to come back up.