vCenter HA configuration fails with "vCenter HA has an invalid configuration"

Products

• VMware vCenter Server

Issue/Introduction

When attempting to configure vCenter High Availability (VCHA), the "Clone virtual machine" task fails.

• The error visible in the vSphere Client for the task is: "An error occurred while quiescing the virtual machine".
• The error visible in the vCenter HA configuration tab is:
  
  
• Analysis of the vmware.log file for the vCenter Server Appliance (VCSA) virtual machine shows a related error: "Couldn't get current disks".
• Symptoms may also include the VCSA's virtual disks showing an "Empty" or "0" capacity in the "Edit Settings" dialog, or incorrect VMDK mappings.

Environment

8.x

Cause

The virtual machine's configuration file (.vmx) is inconsistent with its virtual disk descriptor files (.vmdk). The .vmx file may be pointing to non-existent descriptor files, or the descriptor files themselves may be corrupt or missing.

This state prevents the ESXi host from reading the virtual disk metadata. As the VCHA setup process requires a quiesced snapshot to clone the active node, the operation fails because the hypervisor cannot access or lock the disks. This inconsistency is often caused by a previous failed snapshot or backup operation.

Resolution

Primary Solution: Restore from Backup

• The only supported, safe, and reliable solution is to restore the vCenter Server from your most recent file-based backup.
• Attempting to manually rebuild the disk descriptors is extremely high-risk and, even if successful, will result in data loss, as the snapshot chain is broken.

Secondary Solution (Last Resort / High Risk)

• If no backup exists a remapping from the existing core data disks files

1. Power off the vCenter Server Appliance (VCSA) virtual machine.
2. Log in via SSH to the ESXi host where the VCSA VM is registered.
3. Navigate to the VM's directory on its datastore: 
   

   cd /vmfs/volumes/<Datastore_Name>/<VCSA_VM_Name>/

4. Identify the core data disk files (e.g., <vm_name>-flat.vmdk) and note their exact sizes in bytes.
5. Follow the official procedure to recreate the disk descriptor file(s): (please see Additional information).
   1. This involves using vmkfstools to create a new, temporary disk descriptor and modifying it to point to the existing -flat.vmdk data file.
   2. Ensure the adapter type (e.g., lsilogic) and disk size in the new descriptor match the original configuration.
6. After creating the new, valid descriptor file (e.g., <VCSA_VM_Name>.vmdk), edit the VM's
7. configuration file (.vmx) using a text editor (e.g., vi).
   1. Locate the disk definition lines (e.g., scsi0:0.fileName = "...").
   2. Modify these lines to point to the newly created descriptor file(s).
      1. Example: scsi0:0.fileName = "<VCSA_VM_Name>.vmdk"
   3. Remove any lines referencing old or invalid snapshot files (e.g., <VCSA_VM_Name>-000001.vmdk) that are no longer part of a valid chain.
   4. Save and close the .vmx file.
   5. Reload the VM's configuration on the ESXi host:
      1. Find the VM's ID: 
         

         vim-cmd vmsvc/getallvms | grep <VCSA_VM_Name>

      2. Reload the VM: 
         

         vim-cmd vmsvc/reload <VM_ID>

   6. From the vSphere Client, power on the VCSA VM.
   7. Once vCenter Server is fully operational, retry the VCHA setup task.

Additional Information

Recreating a virtual machine disk (VMDK) descriptor file

Consolidating snapshots in vSphere (Related, as failed consolidations can lead to this issue)