Hardware and RAID configuration:

20 x 4-TB drives configured as RAID 6.

Problem:

Our longtime customer contacted us seeking for a way to bypass all virtualization layers and restore the end-user data directly. If this can be done, it will save about 42 TB of storage space and 6 days supposedly used for staging steps.

Diagnosis:

• All RAID settings are provided. We double-check and found no abnormality. All drives seem to store the most current data.
• The host Ext4 volume stores 42 iSCSI files, approximately 1 TB each. These iSCSI files form a large VMFS volume, where all the virtual machines reside. The main goal is to restore these virtual machine files directly without saving the LUNs.

Solution:

• Run File Scavenger® in special mode in which only the metadata of the 42 iSCSI files are restored.
• Parse the metadata and create a large mapping table for the VMFS volume.
• Modify the File Scavenger® code so it can accept the manually-fed custom volume definition.
• Once the VMFS volume is well defined, run the scan and restore the virtual machine data.

Result:

• 17-TB of virtual machine data is restored. The user-files inside the virtual disks is also checked to ensure data integrity. It is a successful recovery.