Invalid VM entries appear on the source host despite the same VMs successfully migrated and running normally on the destination host

Products

• VMware vSphere ESXi

Issue/Introduction

• vCenter Server/Host Client: The source host inventory shows VMs as "(invalid)".

• Running vm-support -V on the source host may still list the VMs as registered, while the destination host confirms the same VMs are powered on and functional.

• The following log entries from journalctl.log and hostd.log indicate a failure to access the boot media and the subsequent failure to load the VM configuration file (.vmx) because it is locked by the destination host.

/usr/bin/journalctl -b -0 | less

 YYYY-MM-DD-TT vcenter vpxd[6062]: [vim.event.GeneralHostWarningEvent] [warning] Issue detected on ESXi_Source: Bootbank cannot be found at path '/bootbank'

/var/run/log/hostd.log

YYYY-MM-DD-TT In(166) Hostd[2100403]: [Originator@6876 sub=Vmsvc.vm:"/vmfs/volumes/.../vm_name.vmx"] Failed to load virtual machine: N3Vim5Fault12FileNotFound9ExceptionE(Fault cause: vim.fault.FileNotFound
YYYY-MM-DD-TT 798Z In(166) Hostd[2100403]: [Originator@6876 sub=Vmsvc.vm:"/vmfs/volumes/.../vm_name.vmx"] Failed to load VM

YYYY-MM-DD-TT In(166) Hostd[2100403]: [Originator@6876 sub=Vmsvc.vm:"/vmfs/volumes/.../vm_name.vmx"] Failed to load virtual machine: N3Vim5Fault12FileNotFound9ExceptionE(Fault cause: vim.fault.FileNotFound
YYYY-MM-DD-TT 798Z In(166) Hostd[2100403]: [Originator@6876 sub=Vmsvc.vm:"/vmfs/volumes/.../vm_name.vmx"] Failed to load VM N3Vim5Fault12FileNotFound9ExceptionE(Fault cause: vim.fault.FileNotFound
YYYY-MM-DD-TT In() Hostd[XXXX]: VigorOffline_Init: Failed to initialize VIGOR offline: Cannot open file "/vmfs/volumes/.../vm_name.vmx": Device or resource busy. YYY-MM-DDT In(166) Hostd[XXXXX]: State Transition (VM_STATE_INITIALIZING -> VM_STATE_INVALID_LOAD)

 

Environment

VMware vSphere ESXI 8.x

Cause

The issue is caused by a permanent failure of the ESXi boot media (e.g., SD card or USB). Because the /bootbank partition becomes unreachable, the host cannot persist changes to the vmInventory.xml file.

When the VMs are migrated away, the source host attempts to update its local inventory, but the failure of the boot device prevents this change from being saved. Upon a host reboot, the ESXi host loads a stale version of the inventory file. It attempts to register VMs that it no longer owns; since the destination host holds the file lock on the .vmx files, the source host marks them as Invalid.

Resolution

Step 1: Reach out to Hardware vendor to review the host boot media health status (e.g., SD card or USB). 

Step 2: Confirm VM Status: Log into vCenter and verify that the VMs are "Powered On" and healthy on the destination host.

Step 3:  To Unregisters stale entry of VMs :

• Access Source Host: Log into the VMware Host Client directly for the source host (using the host IP/FQDN).
• Locate the virtual machines displaying the Invalid status.
• Unregister VMs: Right-click the invalid virtual machine.
• Select Unregister. (Note: Do not select "Delete from Disk" as this would attempt to delete the actual VM data.)
• Click Yes to confirm the removal from the host inventory.

Additional Information

Virtual machines appear as invalid or orphaned or inaccessible