Specialist work · Synology NAS
Behind that one red banner sit three separate problems: a RAID member that has vanished, LVM metadata that no longer lines up, or a btrfs tree that has stopped making sense. DSM reports the lot in the same words. Each wants a different fix, and the wrong fix costs you data. So until the disks have been read properly, do two things: switch the unit off, and ignore Repair.
Run the problem past an engineer
0800 6890668
| The problem | What's behind it | First move |
|---|---|---|
| DSM is reporting Volume Crashed | One message, three possible culprits: the RAID, LVM or btrfs | Switch it off; leave Repair alone |
| The Storage Pool reads Degraded | A drive has dropped out; the box still serves files, unprotected | Time is no longer on your side |
| The DiskStation will not power up | Nothing has moved; your files live on the drives, not in the chassis | About the easiest case we get |
| The rebuild onto a replacement drive stalled | All that heavy reading exposed a second worn member | Stop the rebuild; clone first |
| Windows wants to format the NAS drives | Normal enough: Windows cannot read a Linux volume | Always decline |
| Shares gone, a ransom note in their place | One of the strains that scans for exposed NAS units | Photograph it, power off, then read our ransomware page |
Every item gets its own case number on the day it lands here. An engineer traces the fault, gives you a plain assessment of what is realistically recoverable, and puts one fixed price in writing. The diagnosis is free, you are under no obligation, and chargeable work starts only when you say go.
Nothing else happens until every disk has a sector-mapped clone. The unit stays powered off throughout, so no rebuild can start itself and the earlier btrfs roots remain exactly as they were.
Everything runs on the clones: slices mapped first, then each mdadm array reassembled, then LVM read out, and last the btrfs or ext4 volume mounted read-only. The sequence is not optional, and we follow it.
RAID, LVM or btrfs — the repair goes onto the rebuilt copy and nowhere near your drives. If the live btrfs tree is beyond saving, we fall back to an earlier sound generation, then open and check every share.
You get the full list of recovered files first, and nothing is charged until you approve it. Data goes back on fresh media, return postage paid, and the job stays open on our side until you confirm every file opens at your end.
This is the detail that settles most cases: put unequal drives in an SHR pool and there is no single array to find. What exists is a set of mdadm arrays built on slices of each disk, tied into one filesystem by LVM. Tackle that in the wrong sequence and you see nothing at all; unpick it properly and a 'crashed' volume is usually complete.
Two 8TB disks, two 4TB, one member down — the DSM rebuild ran overnight and finished on Volume Crashed. Inside were three distinct mdadm arrays: two healthy, the third degraded but usable. We rebuilt them off the box, wound the btrfs tree back a generation, and every share came back intact.
Rarely. All 'Crashed' tells you is that DSM could not put the stack back together: a RAID member missing, LVM metadata broken, or a btrfs tree corrupted. Three separate faults, most of them recoverable, and the warning cannot tell them apart. Leave the box off and nothing is closed off.
No. Take the badge off and SHR is plain Linux underneath: mdadm software RAID over partition slices, LVM pooling those, and btrfs or ext4 laid on top. With clean images of the drives it comes back together on a normal workstation. Trouble starts only when a tool — or an engineer — assumes a single flat array.
Only after each member disk has an image behind it. A rebuild pulls every surviving drive from one end to the other, which is the likeliest single way to lose a second one, and it can overwrite filesystem structures that a recovery still has to read. Take the images first and the rebuild stops being a risk.
Usually some, now and then everything — it depends which strain hit you. NAS boxes exposed to the internet have been a target for years, so we look at snapshots, replication copies and whatever survives in free space. Photograph the ransom note, shut the unit down, then work through the options on the ransomware page.
Switch a failing device on again and you lose a little more. Open a case before you do — the diagnosis is free, whatever it finds.