As a member of a worldwide consortium called the TSC or Technical Services Consortium, we are part of a family of companies that offer many of the same services, but in different geographical areas. Because we are not competing with each other, we have developed some amazing business relationships with folks that would normally be our competitors. This has allowed us to gain knowledge in areas where we were weak and to share knowledge where we are strong. As the only member of the consortium offering data recovery services, we often get called in to help with data loss situations.
Recently one of our TSC partners in the Netherlands, Sprague Europe, was approached by a well-known University in Germany. The University had had a catastrophic failure with a small RAID 5 server and the hard work of hundreds of students was in jeopardy.
Initially the University had contacted a well-respected but smaller data recovery company to help. They were able to image one of the 3 drives used in the RAID 5 array, but unfortunately, all attempts at resurrecting either of the remaining two drives was met with failure.
With RAID 5 and it’s built in redundancy, we only needed to image one of the two non-responsive drives to be able to rebuild the RAID. Both drives had developed light rings on their disks and the heads were physically destroyed by travelling over the damaged areas and had become stuck in the head loader ramps. With each recovery attempt, these rings would increase in depth and width, and soon no recovery would be possible. The smaller data recovery company had exhausted their supply of parts drives and realizing the chances for success were decreasing with each attempt, they advised the University to send it to the largest data recovery company in the world for the best chance. Fortunately they had a branch in Germany and off the case went.
But again, all attempts to recover the data were met with failure and the University was almost ready to concede defeat. But as luck would have it, the head of the University’s Computer Centre was an old colleague with one of our friends at Sprague and in a last ditch effort he called Sprague for advice. Our friend at Sprague suggested Memofix “may” be able to help.
Unlike most data recovery service providers, Memofix does not charge an extra fee for working on hard drives that have been previously worked on by other data recovery companies. In fact we cherish those data recovery cases because it gives us a chance to show just how good we are. Over the years we have recovered approximately 70% of the recoveries that have been previously attempted by our competitors.
When I first saw the request and the summary of what had been attempted, I was skeptical if we could do any better than the two professional data recovery companies that had already tried, especially when one of them was a “world leader” in data recovery and cost did not appear to be a constraint. I spoke with the original data recovery company that tried the 1st attempts. They told me that both defective drives had heads that were bent, distorted, and stuck in the head ramp loaders. They also told me that they had gotten both defective hard drives to function for a very short period of time after replacing the head assembly. Things were looking a little more promising.
When it comes to recovering data, there is one MAJOR factor that influences the chance of recovering data and that is the condition of the disks. If the disks are in good shape, a recovery is usually going to be possible. If the disks are badly scored and ringed, then it will be next to impossible to get a set of heads to function inside the drive without damaging them almost instantaneously. For a disk to get that badly damaged usually means the drive has been running/spinning in a crashed state for a long time. The fact that these drives had their heads stuck in the loading ramps is good news, as they could no longer damage the disks. The second fact that made us hopeful was that the drives were manufactured by Western Digital and Memofix has one of the best Western Digital data recovery techs that we have ever met and we do know a lot of data recovery techs.
So we decided that we may have a chance of successfully recovering at least some of the data and consequently the case was sent to us for a final data recovery attempt.
Our initial evaluation confirmed that we would need to get one of two defective drives imaged. After much scrutiny it was determined that one of the defective drives had shifted disks. When heads are replaced, there is a procedure for adjusting the disks which can sometimes help in getting access to a non-responsive hard drive. I suspect one of the data recovery service providers had attempted this procedure or something similar. This procedure can sometimes make things worse, so we decided to concentrate on the more pristine of the two drives.
Our Western Digital data recovery specialist got right to work. His first task was to get the drive functional enough to read off the drives system area, a normally inaccessible storage area on the drive used for its own housekeeping and the storing of drive specific settings. The system area tells the drive how many heads are present and what settings to use with each head. Without this road-map, recovery may not be possible.
Bent Heads vs Good Heads
We quickly determined that two of the disk surfaces had physical damage and we would worry about those surfaces last. Modifying the system area allowed us to configure the drive to ignore the 2 damaged heads/surfaces. We then physically removed the 2 heads (corresponding to the damaged surfaces) from a good parts head stack. This would prevent further damage to those disk surfaces as well as preventing read interference from the heads going over the damaged areas. We then used this head assembly to get the drive to initialize and we started to image off the good surfaces. After approximately 20 hours of imaging, the heads assembly became too contaminated with crash debris to function and we had to modify a new set of heads to continue. After two days we had managed to image 6 of the 8 disk surfaces. We were now ready to go after the two damaged disk surfaces.
At this stage, our greatest danger was the inherent risk of developing new crashes & destroying our chances of imaging off the last two disk surfaces. Consequently, the drive was imaged only while it could be monitored by a dedicated human watching and listening for any potential new crashes. Imaging of these last two surfaces took almost a week and in the process we consumed five sets of heads and still had some areas where nothing could be read because of the damage to the hard drives’ disks.
With the best drive images we could obtain, we began to analyze what we had. Once we determined the correct RAID 5 configuration settings we were able to use the RAID’s parity to virtually rebuild the missing disk and to de-stripe the 3 drive RAID 5 volume onto a single drive. Then it was a matter of repairing some minor file system structure damage, mounting up the volume, and copying the data to a new target hard drive.
In the end, we were able to recover over 99% of all the students’ data. But there wouldn’t have been any data recovery had it not been for one individual at the University refusing to take NO for an answer.
Consequently, I think the real takeaway point, has to be the fact that not all data recovery companies have the same abilities and just because the world’s largest data recovery company says it can’t be done, doesn’t mean Memofix, the world’s most determined data recovery company, can’t prove them wrong.
We are Memofix, and YES we can recover your data, even if others can’t!