In a datacenter with at least two hosts, a VMFS datastore spans two LUNs, also called storage extents. You use third-party software to create a snapshot of the VMFS datastore. When the original VMFS datastore becomes corrupted, you mount the snapshot on the first ESXi host in the datacenter. The whole VMFS volume, with both LUNs, becomes available to the first host.
A problem occurs when you mount the snapshot on a second host. When second ESXi host attempts to mount the second LUN, it cannot determine whether to mount the second LUN in the snapshot, or the second LUN that the first ESXi host mounted from the snapshot. The operation fails and displays an error.