Virtual machine power-on failure in NFSv4 datastore after network outage.

Products

• VMware vCenter Server

Issue/Introduction

• The virtual machine fails to power on with the below error:

  Failed to start the virtual machine (error -18). Cannot open the configuration file /vmfs/volumes/<datastore_name>/<vm_name>/<vm_name.vmx>

• The power-on attempt may also time out with the below error :

Module VMMon power on failed. Could not power on virtual machine: Timeout. Failed to initialize swap file '/vmfs/volumes/<datastore_name>/<vm_name>/<vm_name.vswp> : Timeout File system specific implementation of OpenFile[file] failed File system specific implementation of CreateFile[file] failed

• The following message is observed in the vmkernel.log and persists after VM re-registration:

     NFS41: NFS41FileOpOpenFile:####: Open of obj 0x############ name <VM_NAME>-flat.vmdk (fhID 0x########) failed: Busy

• The impact is isolated to the specific virtual machine, while all other virtual machines residing on the same host and datastore continue to operate without disruption.

Environment

VMware vSphere ESXi 8.x

Cause

The issue originates from orphaned locks on the .vmx.lck or .vswp files. If the ESXi host fails to clean up file handles correctly during a network outage, the storage layer continues to flag the files as 'Busy' or 'Already exists'. This prevents the initialization of the VM configuration during the power-on sequence.

Resolution

1. Reboot the ESXi host where the virtual machine was previously registered or running. This often releases the stale file handles.

2. If the issue persists after a host reboot, contact the storage vendor to investigate the issue further.