Why Data Recovery Companies Need All Your RAID Hard Drives

We are often asked by clients who require our RAID recovery service whether they should send us all the hard drives from their broken RAID array or just the faulty one(s). The answer is that we always prefer to have all the drives, even the ones that don’t have a problem.

While this may sound counter intuitive (why would a data recovery company want hard drives that are not broken?) the answer is quite straightforward: due to the complex way RAID stores data on hard drives, it’s necessary to have all the hard drives from a failed system in order to successfully complete the data recovery operation.

As it’s frequently not possible to actually repair a damaged hard drive (see our blog post from the 22nd April here), the emphasis of our effort is focused on recovery rather than repair. By having all the hard drives that make up the RAID we are able to fully recover all the data from it.

Sending us one drive from a broken RAID for repair or cloning the drive’s data onto another so that the new disk can then be reintroduced into the array is fraught with danger and should be avoided: see our Top 5 Things To Avoid When You Have RAID or Server Failure

How data is organised on a 3 disk RAID 5

The diagram to your right shows how a file is saved to a server running a RAID 5 setup. RAID 5 isn’t usually proprietary, so this diagram applies to NAS hard drives running as RAID 5 as well a the larger data centre RAID 5 servers built by Dell and HP. At the top of the diagram is a typical file, which in this case has been split into 6 blocks. These blocks are saved in distributed form across all the hard drives in the RAID 5 set, in this case three. The Parity block is what makes RAID 5 such a good selling point as it means that one of the three disks can fail without any data being lost, as the Parity block enables the data from a failed disk can be reconstructed from information contained in the Parity block.

Using this 3 disk RAID 5 as an example, it is necessary for two of the drives to fail before the RAID stops working (because the Parity block allows the RAID to continue to run with just two drives). If only the two failed hard drives were sent to us, it would be difficult for us to obtain any type of recovery because we wouldn’t have the working drive for reference – think of the recovery as a complex jigsaw puzzle, with the working drive being the picture of the puzzle on the box. This is why it’s always preferable to send all the hard drives in to us when dealing with RAID problems.

For more about Data Clinic’s RAID data recovery services click here for RAID, and here for NAS