NDR / Every kind of NAS / Four-bay RAID 5 and 6

RAID 5 · RAID 6 · SHR · SHR-2 · X-RAID · TRAID

Four-bay NAS in RAID 5 or 6. One drive can fail. It is the second one, during the rebuild, that brings people here.

A four-bay NAS in RAID 5 stripes the data across all four drives with a parity block on each stripe, so that any one drive can fail and be rebuilt from the other three. RAID 6 and SHR-2 give two parity blocks and survive two. Synology's SHR, Netgear's X-RAID and TerraMaster's TRAID are the same arithmetic arranged to cope with drives of different sizes, layered in mdadm arrays joined by LVM. All of them share one weakness: the rebuild reads every sector of every survivor, and four drives bought together are four drives that age together. The second failure during the rebuild is the commonest four-bay job there is. A four-bay set is £500 + VAT upwards after the free look, fixed in writing, 5–10 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 it down. Label each drive with its bay number. Do not rebuild, repair, initialise, reinstall or reset. A rebuild reads every sector of every surviving drive, and on a set with a second weak member it finishes what the first failure began. Nothing on the drives gets worse while the unit is off.

Four-bay symptoms, and how long each gives you.

Not listed? Describe it on the form →

One drive red; the pool says degradedRedundancy is gone. Copy what matters, then power down. Do not rebuild.Do not delay

The rebuild stopped, or a second drive dropped during itThe classic. Power down now; the set is recoverable from images.Do not delay

The pool says crashed and the volume has goneTwo members lost, or the metadata damaged. Power down.Do not delay

A power cut mid-rebuildA half-written rebuild. Do not restart it.Do not delay

Drives moved between bays, and the unit wants to initialise themSay no. The order is read from the metadata.Do not delay

Btrfs reports checksum errors after the rebuildThe rebuild used a survivor's bad sectors. Image before anything is repaired.Do not delay

Packing it and posting it: power the unit down, write the bay number on each drive with a marker before it comes out, and send every drive from the set, including any SSD cache. Each drive travels in an anti-static bag inside its own padding, in a box with nothing able to move; the unit itself comes too if it holds an encryption key or a proprietary layout, and its power supply with it. Insure the parcel for what the files are worth rather than the price of the drives, and use a tracked service. 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.

How it is laid out, and what fails.

RAID 5 and RAID 6Data and parity striped across every member. One failure is survivable; two in RAID 5 is not, and the second usually comes during the rebuild. On the bench every member is imaged and the parity arithmetic is run on the images, which cannot fail half-way.

SHR and SHR-2Synology slices drives of different sizes into equal pieces, builds an mdadm array per slice size and joins the arrays with LVM. Two or three layers where a plain RAID 5 has one, each rebuilt in turn from the images. Synology's own note: Auto Repair is not supported on SHR pools.

X-RAID2 and TRAIDNetgear and TerraMaster do the same trick under their own names. A failed X-RAID expansion or a TRAID+ conversion that stopped leaves two geometries on the drives, and the reshape point is found before either is assembled.

QNAP poolsmdadm under LVM under a thin pool under ext4, or ZFS on QuTS hero. A thin volume scatters its blocks across the pool, so the thin-pool metadata has to survive or be rebuilt before the file system can.

What it says, and what it means.

Describe yours to us →

What you see	The usual reason for it	Where that leaves you
Degraded, volume readable	One member lost	Copy critical files; power down; no rebuild
Rebuild stalled at n per cent	A survivor's unrecoverable read errors	Every member imaged, weak areas last; set rebuilt from images
Crashed; volume missing	Two members gone, or metadata damaged	Order, chunk and parity read from the superblocks; assembled in software
Volume mounts read-only with Btrfs errors	A rebuild that used bad sectors	Imaged before repair; repaired on the virtual volume
Drives shown as new after a move	Bays swapped; the unit lost its map	Do not initialise; metadata decides the order
Pool full; thin volume unmounted (QNAP)	Thin pool out of space	Thin metadata rebuilt from the images

From the box 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 drive the day it is opened, and an engineer settles what has actually happened before anything spins. The unit is examined on its own; each drive is assessed on our own equipment, never in a NAS that will try to rebuild. 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 repairs, no resets

The unit, and the drives, apart

A four-bay set arrives as four drives and a unit, and each drive is read on its own before anything is assembled. The metadata on each gives the role, the order, the chunk size, the parity layout and the event count that says which member dropped out first; the weak or failed member goes to the clean bench.

Unit and disks assessed separatelyFailing members to the clean bench

Every member imaged, once

Nothing is worked on live and nothing is rebuilt on the originals. Every drive in the set is imaged sector by sector, the weak ones head by head with the bad areas last, and the SSD cache with them where there is one. From then on the originals are not touched again.

Every drive, including the cacheWeak areas last

The set put back together, from the images

Order, chunk size, parity rotation and the reshape point are read from the drives' own metadata and the array is reassembled in software: mdadm and LVM under SHR, X-RAID and QTS pools; a ZFS pool imported read-only; Drobo's BeyondRAID zones reconstructed. The file system, Btrfs, ext4, XFS or ZFS, is repaired on the virtual volume, and encrypted volumes are unlocked there with the key you supply. Never on the originals.

Rebuilt in software, from imagesThe file system repaired on the virtual volume

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

The rebuild is the most dangerous button on the box. It reads every sector of every survivor; on four drives of the same age, that is where the second failure is found.
The bench never rebuilds. Each member is imaged, the weak one slowly with its bad areas last, and the parity is run on the images. A rebuild on images cannot stall.
Label the bays before a drive comes out. The metadata usually gives the order, but a set whose drives have been swapped between bays and initialised is a harder job.
Send the unit too. The controller's own configuration lives on its board, and on some units so does the encryption key or the cache.

How much comes back: from a RAID 5 that lost a second drive during the rebuild and was powered down, usually nearly all of it, because the dropped drive still holds the stripes the rebuild had not reached. From one that was rebuilt again, reset and reinstalled, less.

One job, followed all the way through.

UK · NDR-2026-0540JOB LOGGED ✓

A Synology DS920+ in SHR with four 4TB drives, drive 2 failed, then drive 4 at 38 per cent of the rebuild

The owner powered down when the rebuild stopped, which is why this was a good job. Drive 4 had unrecoverable read errors and was imaged slowly on the bench with its bad areas last; the original drive 2 was read for the stripes the rebuild had not reached; the two healthy drives were imaged end to end. The SHR arrays were assembled from the images, the LVM volume mapped, and the Btrfs file system mounted read-only from the virtual volume. Every folder came back.

99.98% of the volume recovered7 days here, and back by post

Illustrative example — replace with a genuine case

What helps, and what harms.

Do this much first

Power it down when the rebuild stalls, or before it starts
Write each bay number on its drive
Send every drive, the unit and its power supply
Tell us the RAID mode, the OS version and what was clicked

What sets us back

Rebuilding, or rebuilding again
Fitting a new drive to a degraded set before the data is safe
Clicking Repair, Remove or Manage
Swapping drives between bays, or moving them to another unit

Questions answered before you commit.

A second drive failed during the rebuild. Is everything gone?

Usually not. The drive that dropped out first still holds the stripes the rebuild had not reached, and the drive that failed during the rebuild usually reads on the bench with its bad areas taken last. Both are imaged and the set is rebuilt from the images.

Is SHR harder to recover than RAID 5?

It has more layers, two or three mdadm arrays and an LVM map, and each is rebuilt in turn. It is no less recoverable, and the metadata is well behaved.

Should I try the maker's repair first?

Only if every surviving drive is healthy and you have a backup. On a set whose survivors have been through the same years, the repair reads the bad sectors that end it.

What does it cost?

A four-bay set is £500 + VAT upwards after the free look, fixed in writing, and eight bays or more start at £1,250 + VAT.

How long does it take?

5–10 days at the bench for most four-bay sets.

Begin here if yours is doing the same thing.

Synology volume crashed →QNAP volume unmounted →ReadyNAS volume dead →Synology →QNAP →Netgear ReadyNAS →

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 unit off and the drives in their bays.

0800 6890668

Price my job Packing and posting

Every kind of NAS

Manufacturers

Whatever it is saying now

What it costs, and what follows