Serial attached SCSI devices have been a high speed blessing to the SAN, NAS, and Enterprise Storage Array class of systems. These devices offer a clear solution for any storage requirement due to capacity and configuration. Whether it is Fiber Channel mounted VMware or iSCSI configured raw data access these devices are undoubtedly a smart choice for any enterprise.
With that being said when these systems go down due to multiple hard drive failures it is up to the SAN Data Recovery expert to extract the data in its original form and allow the client to bring their system online.
There are numerous anomalies that can occur to cause a RAID to go down and become unusable. One of the single most prevalent is the condition where one drive has dropped out of the array unnoticed and then a second drive drops out at a later date and brings a RAID 5 down. The Hitachi SMS 100 offers a RAID 6 solution which allows for two drives to drop out of the array. I have in my own experience recovered data from both the RAID 5 and RAID 6 configurations where multiple drives have dropped out of the array. The recovery process is normally very straight forward. The technician images all the drives in the array, acquires the RAID order and configuration, calculates the RAID data offset and de-stripes the array. There are many times when once the RAID configuration has been discovered the array can be software enabled virtually and then the data may be copied onto more reliable media. This litany of steps has been tried and true for many years however recently a small fly in the ointment has appeared.
In order to facilitate a more robust RAID recovery environment an eight byte ECC/Block ID has been added to the end of the standard 512 byte sector. This sector expansion has been implemented at the hardware and firmware level and offers a distinct challenge to the data recovery technician. Outside of its native environment the Hitachi and Fujitsu SAS models configured in this manner cannot be recognized by many of the conventional SAS RAID cards. DTI Data found one card that was inexpensive and would read the data from the aforementioned drives. When using this card with conventional imaging software it was found that the eight byte check byte was being pulled as well as the 512 bytes of data. In addition, after every 1024 sectors imaged the data would skip some data and then start at another sector. Initially it was thought we could strip the check bytes off but the skipping and loss of data at 1024 sectors made that impossible. It was becoming apparent that another step would be necessary in order to build a clean and useful image.
Employing a different piece of software, and using a specific set of parameters for that software we were able to ignore the check bytes. DTI built a clean image for each drive in the array. It was then a simple task to take all of the images, find the active drives in the array and build a clean data set.
This particular situation showed up on a different array type with a friend of mine in Europe. The client had sent the array to three other SAN data recovery companies with no answer to the check byte problem. Although his solution was different the 520 byte extended sector became the key to his recovery as well. DTI data is always on the bleeding edge of data recovery methods. To help ensure your valuable data is given the best chance for recovery the experts at DTI Data Recovery are ready to offer our assistance.