Existing VMFS datastores randomly disappear or unmount from specific ESXi hosts in a cluster.
In the vSphere Client, the storage device (LUN) is visible under Storage Devices, but the Datastore column shows "Not Consumed."
Attempts to mount the datastore fail, or the option is greyed out.
Running esxcfg-volume -l returns no output (no mountable volumes found).
VMware vSphere ESXi
This issue is caused by inconsistent LUN presentation (LUN ID mismatch).
For shared storage to function correctly, the storage array must present the same physical volume (NAA ID) using the identical LUN ID (Logical Unit Number) to every host in the cluster.
If Host A sees a volume as LUN 10 and Host B sees the same volume as LUN 11:
ESXi detects the volume signature but notices the LUN ID differs from the cluster configuration or its own internal history.
ESXi identifies the mismatch as a potential "snapshot" or clone to prevent data corruption.
The host blocks the automatic mount of the volume.
The fix must be applied on the Storage Array side. Do not attempt to force-mount the snapshot on ESXi, as this does not resolve the underlying configuration violation.
Identify the Correct ID: Determine which LUN ID is used by the majority of the cluster (or the standard defined by your storage policy).
Unmap (Storage Side): On the Storage Array management console, unmap the LUN from the affected ESXi host(s) that have the incorrect LUN ID.
Remap (Storage Side): Present the LUN back to those hosts, explicitly assigning the correct LUN ID (matching the rest of the cluster).
Rescan (ESXi Side): On the affected ESXi host, perform a storage rescan to discover the corrected path:esxcli storage core adapter rescan --all
esxcli storage filesystem list