VMFS datastore inaccessible after accidental ESXi installation on shared LUN

Products

• VMware vSphere ESXi

Issue/Introduction

• A shared storage LUN previously containing a VMFS datastore becomes inaccessible after a host deployment or reboot.

• The datastore no longer mounts to any ESXi hosts in the cluster.

• All virtual machines residing on the datastore appear as orphaned or inaccessible.

Environment

• VMware vSphere ESXi 7.x

• VMware vSphere ESXi 8.x

• VMware vSphere ESXi 9.x

Cause

A shared LUN was inadvertently selected as the target disk during an ESXi host installation.

This process overwrites the primary partition table, replacing the single VMFS partition with a standard ESXi system partition scheme (e.g., systemPartition, vmfsl, linuxNative). As a result, the structural VMFS metadata (LVM and FDC headers) is physically overwritten by the new operating system files, destroying the filesystem integrity.

Resolution

1. Immediately disconnect or unmap the affected LUN from the host where the accidental installation occurred. This prevents further ESXi system logs, traces, or background I/O from overwriting any remaining raw data blocks.
   
   
2. Contact technical support.
3. If support advises the necessity to further attempt data retrieval, perform one of the following steps:

• • Restore the affected virtual machines from a known good backup.

  • Consult with your storage array vendor to determine if a volume-level snapshot or immutable snapshot (e.g., SafeMode) exists that predates the OS installation.

  • If no backups or array snapshots are available, engage a professional data recovery service. These specialists have tools to locate orphaned data fragments in the unallocated space of the new partitions.