We recently recovered data from two S-ATA disks removed from a Dell XPS desktop. The two Seagate Barracuda (ST31500341AS) disks were configured in RAID 0 mode. Instead of using a hardware RAID controller such as PERC or Adaptec card, the host system was using RAID BIOS. This type of RAID utilises the BIOS along with Intel chipset software (installed on the operating system) to manage the RAID.
Our hard disk diagnostic tests revealed that disk # 0 was in rude health. However, disk # 1 had defective read/write heads #2 and #3. The disk heads are a crucial component of an electro-mechanical hard disk as they read the data from the disk platters. When two or more heads fail, the only way to recover all the data is to remove the head-disk assembly and replace it with a part of the exact same specification. Near-matchs will not suffice. In this case, we already had a donor HDA in stock.
We opened up the faulty Seagate 1.5TB S-ATA disk in our clean-room and used a device known as a “head-comb” to carefully remove the head disk assembly from the disk chassis. The new head-disk assembly was delicately inserted and was then secured using a Torx screwdriver. Just the right amount of torque pressure needs to be applied – too much or too little can result in the head-disk assembly mis-aligning with the platters. The disk then had to be slowly imaged (onto another disk) which would be used in the next stage of the recovery process. It is always preferable to work with disk images as opposed to the original disks to maintain data integrity.
The RAID rebuild process
With disk images of disk 0 and disk 1 attached to our RAID data recovery system, we started the rebuild process. After determining the block order and block size and other parameters, we inputted these into our RAID data recovery system. The rebuild process started and 6.5 hours later and we had a complete and fully mountable NTFS volume. Using random sampling, we checked the integrity of files. Word, Excel, PDFs and Jpegs all opened up perfectly.
Notes on this case
Dell brought out their XPS range of desktops aimed fairly and squarely targeted at the consumer market. A substantial number of the mid-to-high-end models in this range (from around 2008 to 2014) came with RAID 0 (in a two-disk configuration) enabled by default. So, while some owners are (understandably) comforted by the fact that their system has two disks instead of a paltry one, not all realise these disks are joined at the hip using RAID 0. Dell really should have put a health warning sticker on these systems. “Dear User, You have two disks inside your system in RAID 0 configuration. This is for a larger storage volume and should not be considered a backup…”