Stuck Storage vMotion Task and Duplicate VM Processes

Products

• VMware vCenter Server

Issue/Introduction

• When a vCenter Server instance goes down or becomes unavailable during an active Storage vMotion task, the affected Virtual Machine exhibit the following behavior:
  • The VM becomes unresponsive to management tasks.
  • The VM console is blank or does not display a login prompt.
  • Running the command esxcli vm process list on the ESXi host shows two distinct processes for the same VM.
  • The VM may appear "orphaned" or show incorrect status in the vSphere Client once vCenter is restored.

Environment

• VMware vSphere ESXi

Cause

The source VM process and the destination "shadow" process both remain active on the ESXi host(s), causing a lock conflict and management plane instability for that specific object.

 

Resolution

• Identify and terminate the "stale" VM process while preserving the instance associated with the current datastore.
  

  1. Identify the active Datastore: Log in to the vSphere Client (or check the VM configuration file via CLI) to determine which datastore and .vmdk the VM is currently expected to be using.

     • Navigate to: VM > Edit Settings > Hard disk.

  2. Locate duplicate processes: Open an SSH session to the ESXi host where the VM is registered and run:

     esxcli vm process list

  3. Differentiate the instances: Review the output for the affected VM name. Compare the Config File path or backing information for both entries.

     • The "Good" instance is the one actively using the datastore/path identified in Step 1. Do not terminate this.

     • The "Stale" instance is the redundant entry (typically pointing to the source or a temporary migration path).

  4. Capture the World ID: Note the World ID of the stale instance identified in Step 3.

  5. Terminate the stale process: Run the following command to kill the redundant process:

     esxcli vm process kill -t force -w <world-id>
     

  6. Validation: Verify the VM is now responsive in vCenter. If the VM remains in an inconsistent state, a restart of the hostd and vpxa agents on the ESXi host may be required.

   

Additional Information

If the VM fails to respond or the duplicate processes persist after following these steps, please contact Broadcom Support.