The generic tools initially deployed during data recovery procedures are by nature read-only. Does this mean that a data recovery practitioner using these tools cannot in the process of attempting to recover data from a faulty RAID system, damage the very data he or she is trying to preserve?
In the overwhelming majority of cases no damage to data stored on RAID configured hard disk drives is incurred during the recovery process. When undertaking relatively large volumes of read instructions from large RAID configured storage systems however, mechanical and electronic components can become stressed leading to malfunction that can affect the read process.
RAID system data recovery.
We took four hard disk drives installed in a failed RAID 5 configured NAS system and connected them individually to an imaging tool. Two of the hard disk drives produced read errors requiring intervention to allow the process to complete. The disk images produced could therefore not be considered as precise duplicates but only as components that could possibly be re-engineered into a working a system.
Re-building the RAID system.
The image hard disk drives were then re-assembled into their RAID 5 system disk order and the specific RAID 5 configuration table/s re-built and commissioned into a working system that could access the stored data volume, stripped across the individual disks.
RAID data recovery and file checking.
The data files available from the system were transferred to an alternative storage medium and the integrity of the files checked by a data recovery technician.
Corrupted Data Files.
A small number of the available files were corrupt and unreadable however the majority of data files and folders were acceptable to the customer. The data recovery work in terms of value, availability and cost was also a viable proposition and the customer elected to manually rebuild and reconfigure the small number of corrupted files.
The Data Recovery technicians dilemma
The potential for component replacement on the original hard disk drives targeted at producing an error free duplicate image involves a dilemma. Further work introducing new components may produce increased read error counts, may improve the quality and integrity of the data recovered or may involve no improvements whatsoever. i.e. the stored data available to be recovered was corrupt in the first place. Ultimately the decision to undertake further costly and speculative component replacement and RAID system re-build work depends on the quantity and value to the customer of the data read exhibiting as corrupt.