fault.VmfsAlreadyMounted error when resignaturing a snapshot LUN

Products

• VMware vCenter Server
• VMware vSphere ESXi

Issue/Introduction

When resignaturing a snapshot LUN that is currently active/mounted using the vSphere UI or Resolve API, you experience the following:

• The resignature fails.
• In the vpxd.log, you see the error:
  
  [VpxLRO] -- ERROR task-1640 -- host-9 -- vim.host.DatastoreSystem.resignatureUnresolvedVmfsVolume: vim.fault.VmfsAlreadyMounted::
   
• In the vSphere Client event log, you see entries similar to:
  
  fault.VmfsAlreadyMounted.summary

Environment

VMware vCenter Server 6.x

VMware vCenter Server 7.x

VMware vCenter Server 8.x
VMware vSphere ESXi 6.x

VMware vSphere ESXi 7.x
VMware vSphere ESXi 8.x

Cause

This issue occurs when:

• You attempt to re-signature a snapshot LUN which is already mounted/online. 
• The hosts in the DR site don't have the identical copies of the same LUNs/volumes .   

This is by design and these restrictions are enforced by the vSphere client UI and the query/resolve API to avoid any chances of modifying the VMFS signature while the volume is actively in use. 
 

Resolution

To resolve this issue, perform one or both of these options:

• Unmount and detach the active/online volume prior to resignaturing the snapshot LUN. 
• Ensure that the hosts on the DR site have only 1 copy of each snapshot LUN and avoid multiple copies/presentations.  

Additional Information

Once a VMFS backed LUN is detected as a snapshot, there are two options for mounting/interacting with it: 

•  By force mounting the snapshot LUN, the LUN is enforced to register itself with the existing VMFS signature or UUID reference. 
•  By resignaturing the snapshot LUN with this option, the volume is resignatured and essentially provides it with a new UUID reference. 

The first option, force mounting, can be used when the original volume is no longer visible as a datastore. This approach generally keeps the VMs on the datastore intact, though some may appear as invalid and require re-registration.

With the second option, resignaturing, a completely new volume is created. As a result, the VMs on this datastore must be re-registered because the path to the *.vmx file now points to an updated UUID reference.

In order to proceed with the resignaturing option, ensure that the volume in question is not already mounted/online on other hosts. It is by design that when a snapshot LUN is mounted to a host, the vCenter Server prevents the re-signature of the same LUN.

For more information, see Troubleshooting LUNs detected as snapshot LUNs in vSphere