Devices · RAID arrays & servers
Most arrays have enough life in them to make it through the first failed drive. However, what kills them is what follows that initial failure: the long process of rebuilding a RAID 5 array, which can exhaust whatever remains of a weak survivor; a previously failed drive being forced back online; drives being exchanged between their respective bays. RAID 0 has no "cushion" whatsoever: when one drive fails, you lose access to every single file. We are working solely from image copies, so once your bare disks, labelled by bay, arrive from Exeter, the damage stops. First get a quote, either via telephone or by filling out the form, then post us those bare disks. Don't send the server or its enclosure. If you would like to simply drop off your bare disks, they may be taken directly to our Bristol intake laboratory, Monday to Friday, 9am to 5:30pm, with no appointment needed.
Talk it over with an engineer
0800 6890668
| What it says | What it usually means | Step one |
|---|---|---|
| Foreign Configuration Found (Dell PERC) | When records from the card are compared against the data stored within the disks' own metadata, they do not agree any more. | Neither Import nor Clear: take an image |
| 1786 — Drive Array Recovery Needed (HP) | Parity cover has been lost: either a rebuild is waiting, or it has exited half-way. | Shut down and take an image |
| 1784 — Drive Array Drive Failure (HP) | A member disk has failed in the array. | Do not replace drives based solely on failure symptoms. |
| 1788 — Drives Improperly Attached (HP) | The drives were installed into the wrong bays. | Order matters. Stop what you're doing now. |
| Virtual Drive is Degraded / Offline | The logical volume has lost its redundancy (or is no longer accessible at all). | Power off the entire array. |
| 1720 — SMART drive detects imminent failure (HP) | One of your drives still works but is failing and will soon fail completely. | Get it copied as an image today |
Your device will be logged under its own case number on the day it comes into our lab. The engineer who works on your device identifies the problem with it, clearly states what we think has a good chance of coming back and provides you with one fixed quote. There are no charges for this assessment; there is no obligation for you to do anything; and paid work does not begin until you authorise us to start.
Each drive, including all drives that were rejected by the card, is read on specific purpose-built imaging rigs as part of this process. Every subsequent operation occurs using these copies, and no writes are ever returned to your original drives.
The stripe width, disk drive order and the way parity rotates are determined from patterns within the data. The set is then stitched together across the images on lab machines, with your controller kept out of it, and a rebuild is never performed.
After the rebuilt set is created, we put the file systems, virtual machine (VM) containers and databases on it straight, then run all of these through a complete inventory before any of it leaves the lab.
You have access to a complete list of every item recovered from your device, which you approve prior to any fees being applied. The files will be returned via new media (brand new) with the return postage paid by us, and the case remains open for as long as it takes to confirm that all the files open properly on your own machine.
Why does RAID 5 get so much flak? Because the makers of consumer SATA drives say you can expect about one unreadable sector per 1014 bits of reading (about one every 12.5TB). When rebuilding a big single-parity set, each remaining "survivor" will have to send back more than this amount in one continuous read; this is exactly where rebuild failures occur. The number is written into the specs, not folklore, but it's also a living debate as to how much of a difference it makes in real-world use.
On a Friday, during a mid-array rebuild, one additional drive failed and all of the production data stored on the server became stuck. All four drives were imaged; with three complete images of the array and part of a fourth image, we were able to build the array back up and, by Monday, get the factory operational again.
Yes, get a quote first (by phone or with the form), then send all drives in, labelled by bay. It's possible that even a 'failed' drive holds the most current copy of certain stripes. Therefore, when we rebuild, we use the best-read source for each block across all of the drives in the set.
No. All three (disk order, parity rotation and block size) can be recovered from the patterns that exist within your data. This type of analysis is a common occurrence and is not simply guesswork.
Yes. After an array is rebuilt, the focus will immediately turn to the VMDK and VHDX files along with the databases stored on those drives. Each of these will then be checked for mounting capabilities (not merely whether they are present).
Jobs involving multiple drives typically require anywhere from four to seven working days to complete. If there is no business operation due to this outage, the work can continue uninterrupted; please let us know this when you ring and we'll adjust the intake process accordingly.
Every time a failing hardware component is powered back up, there is additional lost data. First open a case with us: the inspection (look over) is always at no cost, whichever way it goes.