RAR / Every level and layout / RAID 6

RAID 6 · P and Q · Reed-Solomon · MD1200 · MSA · RAID-DP in outline

RAID 6 data recovery. Two parity blocks, two failures tolerated, and a third that is usually a read error nobody knew about.

RAID 6 adds a second parity block to every stripe, Q, computed by Reed-Solomon arithmetic rather than a plain XOR, so that any two members can be recomputed from the rest. It tolerates two failures and is the sane default for large drives and for shelves of twelve or twenty-four members, which is where it mostly lives: Dell MD1200 and MD1400 shelves, HPE MSA arrays, Synology and QNAP rack units, TrueNAS pools as RAID-Z2. It fails on a third failure, or on two failures plus unreadable sectors on a survivor, and it fails in the same ways as RAID 5 otherwise: a stale member forced online, a configuration cleared, a rebuild started with a survivor already predictive. On the bench it asks one thing more than RAID 5, because P alone does not prove the geometry; Q must be computed and checked before the layout is believed. A RAID 6 of two to four members is £500 + VAT upwards after the free look; sets of five or more, and shelves, from £1,250 + VAT, 10–15 days at the bench.

Free first lookOne fixed figure in writingNo data, no bill on most jobsReturn postage paid

Rather talk it through? An engineer answers the bench line
0800 6890668

Power down. Label every slot before a drive comes out. Do not rebuild, do not force a drive online, do not import or clear a foreign configuration, do not initialise, and do not run chkdsk, fsck, zpool import -F or mdadm --create on the members. A rebuild reads every sector of every survivor and writes to the replacement; each of the commands writes to the drives. On images, all of them are reversible. On the originals, none of them is.

RAID 6 symptoms, and how long each gives you.

Not listed? Describe it on the form →

Partially degraded, one member outStill one failure of margin. Replace it, if every survivor is clean.Everyday job

Degraded, two members outNo margin left. Power down; image before anything rebuilds.Do not delay

Third drive dropped during a rebuildEvery member imaged; the set reassembled from images.Do not delay

Shelf offline; all members foreignAn expander or cable. The drives are fine; do not clear.Do not delay

Rebuild finished with errorsPunctured or skipped stripes. Image before recreating.Do not delay

Consistency check found mismatchesA write hole, or a stale member. Stop the check.Everyday job

Packing it and posting it: power down, photograph the controller's screen or export its log, and write the slot number on each drive with a marker before it comes out of its carrier. Send every member of the set, including the one that failed first; it holds the stripes the rebuild never reached. Each drive travels in an anti-static bag inside its own padding, in a box with nothing able to move. Send the controller if it was replaced or holds an encryption key; otherwise it stays. Insure the parcel for what the data is worth, use a tracked service, and know that the posting address is not printed anywhere on this site; it arrives by email in reply to the form, with a booking sheet to print; the sheet inside the parcel is what matches it to your enquiry when it is opened. Or hand the sealed parcel in at the nearest of ten drop-off points, your name on the outside and the sheet inside; say where you are on the form and it comes by email. We pay the postage home either way. The whole of it is written up on the guide to packing and posting.

What it tolerates, what fails it, and what the bench has to find.

P and QP is the XOR of the data blocks, as in RAID 5. Q is a second syndrome computed over a Galois field, which is what lets two missing blocks be solved. On the bench P confirms the stripe and Q proves the geometry; a layout that satisfies P but not Q is wrong.

PlacementSome controllers rotate P and Q together, some rotate them separately, and some vendors' implementations put them in places the recovery tools' preset lists do not cover. The metadata usually says; where it was cleared, both parities are computed and their positions found.

ShelvesA RAID 6 of twelve or twenty-four members behind one expander vanishes as a set when the expander fails and comes back foreign. Restored with the head off, it usually imports cleanly; cleared, it is a reconstruction. See the shelves page on the drive site.

The third failureTwo members out and a survivor with pending sectors is a set one read from the end. The rebuild reads every sector of every survivor, and the pending ones are where it stops. Image first.

What a reconstruction has to determine.

Describe yours to us →

Unknown	Where it is read from	What goes wrong when it is guessed
Member order	Metadata; failing that, P and Q analysis across candidate orders	Every stripe reads from the wrong disk
Stripe size	Metadata; entropy boundaries	Files interleave
P and Q placement	Metadata; computing both syndromes on the images	P satisfied by a wrong layout; Q not
Data offset	Metadata; member 0's partition table or superblock	The volume start is missed
Stale members	Event counts; syndrome mismatch pattern	Old data assembled into new files

From the parcel arriving to your files going back.

Work we have closed →

Logged the day it lands, and the first look costs nothing Free

A case number goes on the parcel and a number on every member the day it is opened, matched to the slot you wrote on it. Each member goes on the imager its interface needs, never on a controller, and its metadata is read before a sector is: the level, the order, the stripe size, the event counts that say which member is current and which dropped out first. An engineer settles what has happened to the set and how much of it can honestly be read back. Back to you come two things together: a straight note of what is liftable and what is not, plus one figure, fixed and written down. Accept it, or decline and owe us nothing.

Nothing to pay for lookingA single figure, put in writingNo rebuilds, no imports, no initialise

Every member imaged, including the one that failed first

Every drive in the set is imaged sector by sector, weak areas last, on hardware that controls every retry, with a map of what could not be read kept for each. Members with failed heads go to the clean bench first; that is the drive site's work and the same bench does it. The drive that dropped out first is imaged too, because it still holds every stripe written before it dropped, and a rebuild that stalled part-way never reached them.

Sector by sector, weak areas lastThe first-dropped member included

The geometry, from the metadata or from the parity

A RAID 6 is proved twice. P across the members at the same offset should XOR to zero on a consistent stripe; Q, computed over the same blocks, should match its own syndrome. A candidate order and rotation that satisfies P but not Q is rejected, and the stale members show as the stripes where both disagree.

Metadata first, parity secondOrder, stripe, rotation, offset

Assembled in software, and repaired on the virtual volume

The set is put together from the images in software, with nothing written to any of them: the current members in, the stale member used only to fill holes a survivor could not give. The file system is checked and repaired on a copy of the virtual volume, VMFS and CSV volumes opened and the virtual machines' disks extracted, databases repaired where they need it. The originals are not touched again.

Nothing written to the imagesVirtual machines and databases opened

You see the file list before you pay

What was recovered is listed for you first, and only then does a bill exist. Approve the list and it is invoiced; turn it down and it is not — and where nothing has come back, most jobs carry no charge at all. Recovered data travels home on fresh media bought in for your job, with the postage at our end. Your case is not closed until you have opened the files on a machine of your own.

No charge until you accept the figureFresh media, supplied with the job5–10 days at the bench

What arrives most often

Two members out is zero margin, not one. The third failure is usually a read error on a survivor, found by the rebuild.
A shelf that went foreign together lost a cable or an expander. The drives are fine and the configuration is intact; do not clear it.
RAID-DP on NetApp is different arithmetic (row and diagonal parity on dedicated disks under WAFL) and is handled as a NetApp job, in outline on the hosts pages.
Send the drive list first, by slot, with model strings. Twenty-four drives take days to image, and the list shortens everything.

How much comes back: from a RAID 6 that lost a third member during a rebuild, usually all or nearly all of it, with the first-dropped members' stripes filling what the rebuild could not compute.

One job, followed all the way through.

UK · RAR-2026-0602JOB LOGGED ✓

A twelve-drive RAID 6 in an MD1200 behind a PERC H810, two members failed a fortnight apart, a third dropped when the rebuild started, and a virtualisation host's datastore on it

All twelve drives came in labelled by slot. The three dropped members were imaged, one after a head swap; the nine survivors were imaged directly. The DDF metadata gave the order and a 256KB stripe, and both syndromes were computed across the images to prove the P and Q placement before the set was assembled. The VMFS datastore was opened from the virtual array and eight virtual machines' disks extracted and checked.

100% of the datastore recovered12 days here, and back by post

Illustrative example — replace with a genuine case

What helps, and what harms.

Do this much first

Power down when the second member goes
Label every slot; send every member
Send the drive list by slot, with model strings, ahead of the parcel
Tell us the controller and the stripe size if known

What sets us back

Rebuilding with a survivor showing pending sectors
Clearing a foreign configuration on a shelf that vanished together
Deleting and recreating after a rebuild with errors
Letting a consistency check run on a set with a stale member

Questions answered before you commit.

Can a RAID 6 with three failed drives be recovered?

Usually, if the three drives can be imaged. The members that failed first hold the stripes written before they dropped; the set is reassembled from images with those stripes used to fill what the survivors cannot give.

Why is Q checked as well as P?

Because a wrong order or rotation can satisfy the plain XOR of P by accident; Q, computed over a Galois field, does not forgive it. The geometry is believed when both agree.

My whole shelf went foreign at once. Are the drives dead?

Almost certainly not. An expander, a cable or a supply failed. Restore it with the head off and do not clear the configuration; it usually imports cleanly.

What does it cost?

£500 + VAT upwards for a set of two to four after the free look; five or more, and shelves, from £1,250 + VAT, fixed in writing.

How long does it take?

10–15 days at the bench for a shelf; 5–10 days at the bench for a small set.

Begin here if yours is doing the same thing.

Hyper-V data recovery →RAID 50 and RAID 60 →RAID-Z1, Z2 and Z3 →Virtual disk failed or offline →Dell PERC →

Nothing gets worse while it is powered down.

Looking at it is free. Back comes a list of what opened and what did not, together with a single price to finish, set down in writing while you are still free to say no. On most jobs an invoice only follows the data. Until that list reaches you, leave the server off and the drives in their slots.

0800 6890668

Price my job Packing and posting

Every level and layout

Manufacturers

Whatever it is saying now

Servers

What it costs, and what follows