The NAS in the closet has been running quietly for four years. This morning it’s beeping, or the web interface says “Volume Crashed,” or a drive shows as failed and the rebuild won’t complete.
NAS cases are recoverable more often than people expect. They’re also the category where the owner’s next few clicks matter most, because a NAS will cheerfully offer to do the exact thing that destroys the array.
Stop before you rebuild
If your NAS is showing a degraded volume and offering to rebuild onto a replacement drive, understand what a rebuild involves: reading every sector of every remaining drive to reconstruct the missing one.
In a RAID 5 that’s already lost one drive, there’s no redundancy left. If a second drive has bad sectors — and after four years of identical drives from the same batch running the same hours, that’s common — the rebuild hits an unreadable sector, fails, and can drop that second drive from the array too. Now you’re down two drives in a set that tolerates one.
Rebuilds on aging arrays fail regularly. If the data matters more than the uptime, image the drives before rebuilding, not after the rebuild fails.
Never click these: “Initialize,” “Create Volume,” “Repair” on a crashed volume, or anything that offers to set up storage fresh. Any of them can write new metadata over the array configuration that recovery depends on.
Record the drive order before you touch anything
This is the single most useful thing you can do, and it takes two minutes.
Number the bays, pull the drives one at a time, and write the bay number on a sticky note on each drive. Photograph the serial number labels in order.
RAID reconstruction depends on knowing which drive held which position and in what order the data was striped. It can be worked out analytically from the drive contents — but it’s dramatically faster and safer when you just tell me the order. Drives handed over in a shuffled pile turn a straightforward job into a puzzle.
What’s actually inside a Synology or QNAP
Worth knowing, because it explains why these recover well.
Synology’s SHR isn’t a proprietary format — it’s Linux mdadm software RAID with LVM layered on top, and the file system is ext4 or btrfs. QNAP is similar. That means the array can be assembled and read on a Linux system without the original NAS hardware.
Which is the good news: if the NAS unit itself died — power supply, board, fan-related heat death — your data is fine. The drives are the array. A dead enclosure is an inconvenience, not a data loss event. Don’t let anyone tell you that you need an identical replacement unit to get your files back.
Free evaluation. Firm quote before any work begins.
What usually went wrong
One drive failed and nobody noticed. The array ran degraded for months because the email alerts were never configured, then a second drive went. This is the most common NAS case I see, and it’s the argument for actually turning on notifications.
A rebuild was attempted and failed partway. Sometimes recoverable, sometimes messier depending on how much got written.
Power event. A surge or a hard shutdown mid-write leaves the array metadata inconsistent. Often very recoverable.
File system corruption on an otherwise healthy array. All drives fine, btrfs or ext4 structures damaged. Usually the best-case version of this.
Expansion or migration interrupted. Adding a drive or converting RAID levels is a long operation that rewrites layout. Interrupt it and the array is mid-transition, which is genuinely difficult work.
What I can take on
RAID 0, 1, and 5 are on my in-scope list, discussed case by case — and NAS arrays are the most common form they arrive in. The work is imaging each member drive individually, determining the layout parameters, assembling the array virtually from the images, and pulling the file system from that.
Note “from the images.” Nothing gets assembled on the original drives. If a member has bad sectors, that gets handled during imaging with the rest of the array safely captured.
Where I’ll refer out: arrays where multiple drives have mechanical failures needing cleanroom work, and larger enterprise configurations beyond a handful of drives. I’ll tell you at evaluation.
The bottom line
Don’t rebuild, don’t initialize, and don’t reorder the drives. Label the bays, power the unit down, and get the array evaluated as a set. A NAS whose enclosure died is usually a straightforward recovery — a NAS whose owner clicked through three repair prompts sometimes isn’t.
NAS showing a crashed volume? Start a case — tell me the model, how many drives, the RAID level if you know it, and whether a rebuild has already been attempted. That last one changes the approach the most.