Redundancy hides the initial failure (hence most arrays come to us having silently lost more than one drive). Every single disk is imaged read-only, and the geometry will be developed by reading the data off the drives instead of trusting a controller that has already failed. The array is rebuilt in software, so you can’t hurt anything on the original disks. Running a RAID 10 striped mirror? RAID 10 explained describes both what it does survive and what it won’t.
One rebuild will write on the member disks. If there is a problem reading parity cleanly, then wrong data will be written. What was once a whole array can be broken into many fragments from one rebuild which was well intentioned. Shut off power and keep bays in original order.
Almost never done as the docs say. Redundancy hides the first failure. Thus, most arrays show loss greater than 1 at arrival.
Since we know the controller is unreliable, we take nothing it tells us as true.
All members of the array are initially imaged read-only (including any disk that was previously identified as “failed”) to see what information is left. In many cases, a removed member will hold some readable data which could provide a means to recover part or all of the lost data. Once the usable data has been extracted, stripe width, disk sequence, parity rotation and data offset are determined based on the structure of that data and candidates for each are examined using trial and error until a valid configuration is found where the reassembled files appear to be structured correctly rather than simply being noise.
A virtual array of disks is created from image files and the file system rebuilt from the virtual array. When there are unreadable sectors on one member of a stripe, parity information can fill in the missing value(s) as long as it is possible to get a valid stripe back. Only when you have two or more failed disk members (in the same location) will the recovered stripe fail. In all cases nothing will be physically written to the source disk(s). Guide to RAID 1 data recovery & RAID 0 data recovery.
While the levels differ in terms of arithmetic, the methodology doesn’t.
Including JBOD, spanned volumes and vendor-proprietary options like Synology’s SHR and Drobo’s BeyondRAID.
Dell PERC, HP Smart Array, LSI and Broadcom, Adaptec, IBM ServeRAID etc. All others (including those with lost configurations) are supported.
Linux mdadm, Windows dynamic disks and Storage Spaces, ZFS and BSD RAID.
From two-disk mirrors to large enterprise configurations. More members = more images = more analysis = higher price.
Arrays that are unknown in terms of either their level or in which order they are located: exactly what defining the geometry is about.
Priced according to the effort required, as opposed to how much storage space you have available.
The cost of recovering your RAID, NAS and server will start at £500 +VAT (dependent upon diagnosis). A free diagnostic can be carried out over a 48hr period. There is a 50% deposit required for members who require physical drive level work. Only on successful recovery will the balance be paid.
Send all disks to us (including any which are currently listed as “failed”), clearly labelled with their bay number; and please provide information about your controllers, such as models, levels (if you know them), and most importantly: What have you tried so far? If you were in the middle of a rebuild when you stopped it, this will make things different. Full reports: an eight-disk RAID 6 a disk too far, a RAID 0 stripe undone by one disk, a RAID 1 mirror worn thin on both sides, a RAID card that lost track of its array and twelve disks initialised by accident.
To begin your free diagnostic of the array, you will need to send its disks in along with the RAID level, number of disks, and type of controller that is on your array. An engineer will then review the information and provide a written quote before we perform any work on the array.
Please obtain a price quote from us via phone or through our online quote request prior to shipping/dropping off your disks. The first step in recovering your data is receiving all of the array’s disks (labelled sequentially); they need to be removed from the server or enclosure, even if you are dropping them off in person. Carefully package the disks, include your contact information with the shipment, and after we have performed a free diagnostic, we will also provide a written quote as to the total cost of recovery services before commencing any work.
Posting it? We would suggest using a tracked and insured delivery method. Want to drop it off? That’s okay too. Monday to Friday, from 9am to 5:30pm is good; no appointment necessary: simply prepare the item as described above.
Do you need a custom quotation? Please describe what type of storage array is being used (what model), how many disks are configured as one unit (RAID levels etc.) and provide information about the failure on the quotation request form and we can have one of our engineers review it, then send you a customised quotation.
We’ll be in touch shortly. If it’s urgent, call 01865 593000.
Every common RAID level can be recovered through our RAID recovery services and as such two levels require separate consideration. A RAID 0 striped configuration will stripe data across all of the disks in the RAID configuration without providing redundancy; therefore if any disk fails, the entire RAID configuration will fail along with it. However, since each member, including the member that has failed, can be copied using an imaging process and then we re-assemble the stripe set virtually from these images, we find this type to be recoverable with ease when used on the lab bench. A RAID 5 configuration will survive the loss of one drive but provides for a significant amount of danger during the rebuild phase. The rebuild will occur by placing additional load on another already tired drive which may also fail. Therefore, with a degraded RAID 5 you should power down and image the system instead of rebuilding the RAID 5 in-place. RAID 6 and RAID 10 configurations are also treated similarly. Members are first imaged, then logical work is done on the copied images.
Questions we get most often regarding RAID recovery.
Often, yes, but it also depends on what happened with the second hard drive that failed. If the other hard drive was simply removed (an “ejected” member), then much of its contents may likely remain readable. All disks will need to be brought into our lab. Every single member of your array will be copied, and your array will be rebuilt using those copies instead of the original physical devices.
No, if you value your data. Once a rebuild occurs, all members are written to. And when there is damage that causes unreadable parity, incorrect data will be written. Rebuilding is the number one reason why a recoverable RAID array becomes irrecoverable. Shut down power immediately and ship your disks as is.
The answer is yes; all disks in your system (including those previously identified as “failed”) will be included in this process. All levels of RAID other than RAID 1 require access to all disk(s) within the array to rebuild the volume. Even an ejected disk from the RAID will usually provide read-accessible data to the extent it was able to write its portion of the data to disk at some time prior to failure. Mark disks with their respective bays prior to removal.
This is normal and should not cause concern. These parameters (stripe-size, member-order, parity-rotation, offset) are generated based upon analysis of the physical media itself; the controller’s metadata is viewed as a suggestion rather than factual.
Yes. Both of these are also proprietary layouts that require derivation from an image rather than reading it directly off the device, much like any other array that can’t rely on its controller.
The cost of the service starts at £500 plus VAT. A written quote will be provided at no charge after a free diagnostic within 48 hours. If the member requires physical drive level work, then a 50% deposit is required with the balance only payable once the data has been successfully recovered.
There is no fee for diagnosing a problem. Pricing varies based upon a number of tiers starting at £500 for multi-disk arrays; however, in most cases you only pay if we recover your data, whether one or multiple levels, controllers, NAS or servers. Contact us today to get started.