RAID arrays don't usually fail one drive at a time — they fail slowly, then suddenly. We recover from RAID 0, 1, 5, 6, 10, JBOD, and any NAS or server platform you're running. The first rule: stop trying to rebuild the array yourself. The second: call us.
Different RAID levels fail differently and need different recovery strategies. Here's what we work on and what survival looks like for each.
No redundancy. Data is split across multiple drives for speed. If any drive fails, the entire array is offline. We can recover by imaging each drive separately and reconstructing the stripe pattern in software.
Two drives, identical copies. Most forgiving config. If one drive fails, the other still has all data. Both failed? We image whichever has more readable sectors and reconstruct from there.
3+ drives. Survives one drive failure via parity calculation. Survives a second failure during rebuild? No — that's how RAID 5 arrays die. Recoverable in most single-failure cases if you stop the rebuild attempt.
4+ drives. Survives two drive failures. The most common config in modern NAS units. When 3 drives fail, recovery becomes a complex reconstruction job — possible but takes longer.
4+ drives in mirrored pairs. Performance + redundancy. Survives one drive in each pair failing. Recovery from a properly-failed RAID 10 is high success rate; recovery from "two drives in the same pair" is harder.
Just a Bunch Of Disks — or Synology's flexible Hybrid RAID. Each drive holds independent data; no parity. We work these by extracting from each drive separately and reconstructing the volume layout from filesystem metadata.
Usually a power surge, controller failure, or drives from the same manufacturing batch ageing out together. We image each drive individually, then reconstruct the array virtually in software.
Drive failed, customer (or IT team) tried to rebuild, second drive failed during rebuild. The array is now offline and the configuration metadata is half-overwritten. Still recoverable in most cases — but only if we get the drives before further attempts.
The controller (hardware RAID card or NAS motherboard) failed but the drives themselves are healthy. We read each drive, identify the controller's stripe layout from the metadata blocks, and rebuild without the controller.
Synology / QNAP / Drobo / TerraMaster shows the drives as healthy but reports the volume as crashed, degraded, or "not initialized". Usually a filesystem-level corruption. Recoverable by extracting drives, mounting via Linux, and repairing the volume metadata.
Someone clicked the wrong button in the management UI. The actual data blocks are usually still on the drives — only the partition table or RAID metadata was wiped. Recoverable by reading raw blocks and reconstructing the filesystem.
Server got hit by ransomware that encrypted everything reachable. We image the drives, look for unencrypted shadow copies, attempt recovery from any partial backup metadata. Honest note: ransomware recovery is highly variable — sometimes great, sometimes nothing recoverable.
RAID recovery pricing depends on how many drives are in the array, what condition each drive is in, the RAID level, and what state the array was in when work began. Here's the honest framework we use to quote.
DS series (DS218, DS220+, DS923+, DS1522+, etc.), RS rackmount series, all SHR / SHR-2 configurations, btrfs and ext4 volumes.
TS series (TS-453D, TS-664, TS-873A, etc.), TVS towers, QuTS hero (ZFS), traditional QTS (ext4), Thick + Thin volumes.
5N, 5N2, 8D, B810n. Drobo's BeyondRAID is non-standard — recovery requires reverse-engineering their metadata structure. We've done it.
TerraMaster F4/F5 series, ASUSTOR Lockerstor + Drivestor, Buffalo TeraStation + LinkStation. Standard RAID configs — recovery similar to Synology.
Server 2012 / 2016 / 2019 / 2022, Storage Spaces simple/mirror/parity, ReFS and NTFS volumes, clustered storage with Failover Cluster.
Software RAID via mdadm (RAID 0/1/5/6/10), LVM volume groups, ZFS pools (RAIDZ, mirror), ext4 / xfs / btrfs filesystems on top.
Dell PERC, HP Smart Array, LSI / Broadcom MegaRAID, Adaptec, Areca. We extract drives, identify the controller's stripe pattern, and rebuild without the original card.
Apple SoftRAID, Mac Pro internal storage, older Promise Pegasus arrays, current Apple Silicon Mac Pros with internal storage cards.
For most RAID jobs we follow up with a phone call to confirm details (drive count, RAID level, what state the array is in) before quoting in writing.
Diagnostic is same-day for business cases. Imaging starts within 24 hours of drop-off. Total recovery typically runs 1–4 weeks depending on array complexity and failure mode. If you need provisional access to specific files (a critical report, customer database) during recovery, we can prioritize extracting those first — ask during diagnostic.
Call us first. A hot-spare rebuild puts maximum read load on every remaining drive in the array — if any of them are also marginal, the rebuild itself can trigger another failure and turn a recoverable scenario into an unrecoverable one. We can advise on whether a rebuild is safe based on the SMART data of your remaining drives.
If the encryption is volume-level (Synology Encrypted Shared Folders, QNAP encrypted volumes) and you have the password, we can decrypt during recovery. If it's drive-level encryption (BitLocker on Windows Server) and you have the recovery key, same story. Without the key/password, we can sometimes recover raw blocks but cannot decrypt them.
Data only — we don't restore an OS to a recovered server because the original hardware is no longer trustworthy. Your data comes back to you on a new RAID-capable target (drives you provide, or we provide). You rebuild the OS fresh on new hardware and copy the data over.
Yes — for any business case involving sensitive data (customer records, financial data, IP), we sign a standard mutual NDA before diagnosis. If you have your own NDA template, we'll review and sign that instead. Your data stays in our shop the whole time and we don't browse it.
Very common — bring the drives in regardless. We do a free diagnostic to assess the current state. If it's recoverable from where it is, we'll quote it. If it's been pushed past recovery, we'll tell you that for free. Better to know than to keep guessing.
Yes for diagnostic. Recovery itself is bound by physics (drives have to be imaged at their reliable read speed; some failed drives need careful, slow imaging to avoid further damage), so we can't shortcut the recovery work itself. But we can prioritize your case ahead of the queue and start imaging immediately for an emergency surcharge.
For RAID and server recovery, the first step is always a phone call. Tell us your array config, what state it's in, and how urgent it is. We'll tell you right away whether to drive it over, ship it overnight, or sit tight while we walk you through the safest next step.