Data Recovery · October 20, 2026 · ComputerKing Team

RAID Recovery for Hong Kong SMEs: What Not to Do First

RAID is not a backup — it only keeps the system running when individual drives fail. After a RAID failure, the most dangerous thing is often not "doing nothing" but "doing the wrong thing": hitting re

<a id="sec-1"></a>

Key Summary

RAID is not a backup — it only keeps the system running when individual drives fail. After a RAID failure, the most dangerous thing is often not "doing nothing" but "doing the wrong thing": hitting rebuild, initialise or format before understanding the situation can permanently wipe data that could otherwise have been recovered.

Contents

• 1. What Is RAID? RAID 0 / 1 / 5 / 6 / 10 Comparison
• 2. Why Does RAID Fail? Common Causes
• 3. Eight Things Never to Do After a Failure
• 4. Common Misconceptions: Four of the Most Harmful Sentences (Plus Two More)
• 5. Handling a Degraded Array: The Safe Process for a Single-Drive Failure
• 6. When to DIY, When to Call Professional Recovery
• 7. How Is Pricing Calculated? Why Individual Quotes?
• 8. Prevention: So Next Time Is Less Frightening
• FAQ
• Sources

<a id="sec-2"></a>

1. What Is RAID? RAID 0 / 1 / 5 / 6 / 10 Comparison

RAID (Redundant Array of Independent Disks) combines multiple hard drives into one logical storage unit. The trade-offs differ greatly between levels — picking the wrong level only becomes obvious when something fails, and by then it is too late.

Sources: Wikipedia, "RAID"; Liquid Web, "RAID Levels 1, 5, 6, 10"; UGREEN NAS, "RAID Levels NAS Guide".

Two common misunderstandings are worth clearing up. First, more drives in RAID 0 does not mean safer — data is spread across all drives, any single drive failure makes the whole volume inaccessible, and the more drives there are, the higher the volume's average failure rate. Second, adding a third drive to RAID 1 adds another copy, not more space; mixing different-capacity drives in traditional RAID uses the smallest drive as the baseline, and the extra capacity on larger drives sits idle (e.g. 2TB + 3 × 4TB in RAID 5 gives (4−1) × 2TB = 6TB usable). Another common trap: manufacturers label TB in decimal (1TB = 1,000GB) while operating systems show TiB in binary (1TiB = 1,024GiB), so a "4TB" drive shows as smaller in the OS — that is normal.

Software RAID is managed by the operating system — usually cheaper, but potentially slower and adding CPU load. Hardware RAID is managed by a dedicated controller — faster, easier drive replacement, but potentially more expensive. Which suits you depends on workload and budget; there is no one-size-fits-all answer.

<a id="sec-3"></a>

2. Why Does RAID Fail? Common Causes

1. Drives from the same batch ageing together

Drives in an array are usually bought and installed at the same time with similar workloads, so they approach end of life at similar times. The classic RAID 5 nightmare is: the first drive fails, and during the rebuild the second drive fails too, because both drives are equally aged and equally stressed.

2. A second drive failing during rebuild / URE

A rebuild must read every sector of every surviving drive. In a large array built from large drives, that is an enormous amount of reading. A single unreadable sector (URE, Unrecoverable Read Error) can halt a rebuild; depending on the controller, the rebuild aborts or marks the affected stripe as bad and continues.

To be precise: URE specification figures are ceilings, not timetables; most drives perform far beyond spec in practice. In large arrays, the main real-world driver of rebuild failure is correlated mechanical failure, not clean per-bit errors. This article therefore gives no single "X% chance of rebuild failure" figure — different sources use different assumptions to produce different numbers, and no single one can be taken as fact. The qualitative statement is enough: the bigger the drives and the bigger the array, the more the rebuild risk deserves attention.

3. Controller failure / lost configuration

If the RAID controller dies or the configuration metadata is lost, the operating system no longer knows how to assemble the array; the drives themselves are healthy, but the array is gone. Swapping in an identical replacement controller does not always fix it, because the new controller may try to initialise the array instead of reading the existing configuration on the drives. Controller failure has another alarming trait: multiple drives can show as failed at the same time even though they are healthy, confusing troubleshooting — many people attempt a random rebuild at this stage, causing irreversible data damage.

4. Accidental deletion / formatting / file-system damage (logical failure)

Not every RAID failure starts with hardware: file-system damage, accidental formatting, deleted partitions, malware, or damaged RAID metadata can make data inaccessible while the drives remain physically fine. Deleting a partition usually only wipes the identifier (the "map"), not the data itself; only creating a new volume, formatting the space, or continuing to write data will overwrite the files you want back. The key condition: whether deleted files can be recovered depends on the drives not being reused or overwritten after deletion.

5. Power outage / surge / improper shutdown

Power outages and voltage surges are a major cause of RAID damage, especially on arrays whose controllers have no battery backup unit (BBU); a write operation interrupted mid-flight leaves the array logically inconsistent, and the controller may not be able to resolve it correctly after power returns. An outage can also interrupt a rebuild: the rebuild stops, some stripes may have updated parity while others have not, leaving the array in a mixed state.

Other causes include firmware / software update failures (especially on consumer and prosumer NAS, where an incorrect update can damage RAID metadata) and human error — pulling the wrong drive, triggering a wrong rebuild, resetting the whole RAID configuration. According to TechChef's analysis, damage from human error exceeds damage from hardware itself.

<a id="sec-4"></a>

3. Eight Things Never to Do After a Failure

Every item below is consistently raised by multiple recovery companies / guides. Remember: the order itself is the point.

1. Do not trigger / repeat a rebuild before confirming the situation. A rebuild puts enormous read pressure on surviving drives; after a first rebuild fails, do not attempt a second one — stop immediately. A rebuild can overwrite existing blocks and permanently destroy recoverable data when the configuration is unstable or incorrect; focus on data recovery first.

2. Do not initialise / recreate / re-initialise the array. This wipes partition data and configuration, making recovery exponentially harder. If the NAS interface shows an "initialize / format" prompt, do not accept it.

3. Do not format the volume / partition. If Windows pops up "You need to format the disk before you can use it", do not format. A volume showing as RAW usually means severe file-system damage, not destroyed data.

4. Do not run CHKDSK / FSCK / file-system repair tools. These tools are designed for single healthy drives; they write to the volume and can destroy the file-system structures recovery tools rely on, making professional recovery more complex or even impossible.

5. Do not swap drive order / pull drives at random. Drive physical order is critical — shuffling them can damage array metadata and mislead the controller into a destructive rebuild. Pulling the wrong drive is one of the fastest ways to turn a recoverable situation into total data loss.

6. Do not write new data to surviving drives / the failed array. Every write increases overwrite risk. Do not reinstall an OS on a drive connected to the failed RAID.

7. Do not repeatedly force the array back online. Do not rely on the controller alone: if the controller or configuration is the problem, individual member drives may hold the information needed to rebuild the array.

8. Do not keep running when there is an abnormal noise. If a drive makes clicking / grinding noises, power off immediately. Continuing to run a physically damaged drive can worsen the damage. Noise only signals "hard stop" — do not treat it as a diagnosis of one specific fault.

What You Should Do Instead

• Stop the array: no more copying new data, no check-disk tools, no improvised repairs.
• Shut down safely: power off to prevent further writes or drive stress.
• Label every drive: mark each drive by bay number and role to preserve the original order; photograph the drive bays before removing anything.
• Record the configuration: RAID level, controller model, stripe size (if known), each drive's model and serial number, which drives failed, what operations were already performed (rebuild attempts, drive swaps, restarts), the exact error messages, and what the system was doing at the time (e.g. backing up, after a power outage, mid-rebuild).
• Keep all related drives: including failed drives, replacement drives, hot spares, and drives removed during an interrupted rebuild — they may hold different versions of RAID metadata or file data.
• Think about backups first: is there a recent backup? Where is it? Can it be restored to another computer / storage device and checked against the latest records? Before contacting professional recovery or vendor support, prepare the information above — do not rely on a verbal "the NAS is broken".

The general approach of professional recovery (conceptual level, not an operating guide): first clone each drive one by one into a safe copy, then perform a virtual rebuild on the copies (including drive order, stripe size, parity layout and other parameters) and extract the files — never rebuild directly on the original drives. Recovered data should be restored to a different healthy storage device, never written back to the original failed array.

<a id="sec-5"></a>

4. Common Misconceptions: Four of the Most Harmful Sentences (Plus Two More)

Misconception 1: "RAID = backup." Wrong. RAID protects against drive failure; it does not protect against accidental deletion, ransomware, fire, theft, or a second drive failing during rebuild. Simply put: "RAID copies everything instantly, including things you did not want it to copy." RAID is about data availability; backup is about data protection. Real backup is essential (the 3-2-1 rule: 3 copies, 2 different media, 1 offsite).

Misconception 2: "If one drive in RAID 1 fails, nothing happens." Not necessarily. RAID 1 mirroring copies instantly: a deleted or damaged file is deleted or damaged on the mirror copy at the same time; ransomware spreads instantly to all mirrored drives; file-system damage can propagate across all drives in sync. RAID 1 is designed for high availability (preventing downtime from drive failure), not for preventing other forms of data loss.

Misconception 3: "RAID protects you from ransomware." Wrong. Once ransomware encrypts your data, every drive in the array is affected at the same time, and there is no clean copy to restore from.

Misconception 4: "A RAID rebuild is guaranteed to recover." Wrong. Rebuilding an array after a drive failure does not always work as expected, and can even overwrite the remaining data. According to professional recovery lab Tapscape, operator error, not the original hardware failure, is the main cause of permanent RAID data loss; swapping the wrong drive, re-initialising the array, or repeatedly forcing the array online are the most common mistakes.

Two more to dismantle: a RAID 5 drive replaced is not instantly safe — after entering degraded mode, the array runs with zero redundancy, and a rebuild can take many hours, sometimes more than a day for large drives; until the rebuild completes, the array stays degraded and fragile. And "the NAS shows degraded, just press rebuild and it's done" is dangerous thinking — if multiple drives look questionable, get guidance specific to that model before starting a rebuild.

<a id="sec-6"></a>

5. Handling a Degraded Array: The Safe Process for a Single-Drive Failure

The following process applies ONLY when it is confirmed that just one drive has failed, the remaining drives are healthy, and a backup exists. If you are unsure, or the data is too important to risk, stop first, preserve the original state, and seek professional help. Do not read this as "a DIY rebuild always works".

Step 1: Confirm which drive really failed. Do not pull drives by guessing. Open the NAS interface (Synology DSM, QNAP QTS, TrueNAS or similar), go to the storage / HDD management page, which marks the failed drive by bay number and serial number; note both. Before pulling, check the on-screen serial against the label on the drive; if the model supports drive identification, trigger the flashing LED on the failed bay to make sure you pull the right one.

Step 2: Pick the right replacement drive. The replacement drive's capacity must be equal to or larger than the failed drive — never smaller; even if retail labels look identical, actual usable capacity can differ. Use a proper NAS-grade drive rather than a cheap desktop drive; NAS drives have error recovery control and are designed for array vibration and 24/7 operation. Avoid drive-managed SMR (shingled) drives for RAID rebuilds, as they can stall long enough to be kicked out mid-rebuild, failing the whole rebuild. Confirm interface and form factor (SATA / SAS / NVMe, 2.5-inch / 3.5-inch, tray / caddy, and supported sector format), and check the NAS or controller vendor's compatibility list and firmware notes.

Step 3: Swap and rebuild — precautions. If the NAS supports hot-swap, you can swap drives while powered on; if not, shut down cleanly first. Remove the failed drive, install the new drive in the same bay (Synology users: deactivate the drive first, per the next section), wait for the interface to detect it, then start repair / rebuild in the storage manager. Keep the system on stable power during the rebuild, ideally through a tested UPS; do not assume "rebuilding" means "fixed"; watch for new read errors, a second drive failing, overheating, or a stalled percentage; unless the vendor explicitly requires it, avoid firmware updates and unnecessary restarts during the rebuild; avoid heavy workloads during the rebuild.

Synology and QNAP Official Process Highlights

Synology (official Knowledge Center, "The system partition has degraded or crashed. What can I do?"): For a degraded system partition, open Storage Manager → Overview → click the Repair link, and the system will start repairing the system partition — wait for it to finish; then go to the HDD/SSD page, and if a drive still shows System Partition Failed status, it may be a failed drive needing replacement. When replacing a failed drive in a storage pool, first deactivate the drive, then replace it, then repair the storage pool or system partition as shown on the Overview page. Multiple failed drives must be handled one at a time: the official footnote states that if there are multiple defective drives, you must handle them one at a time, completing one drive's repair process before moving to the next. For a crashed system partition, the official steps are to back up data first, then use Mode 2 to reset the NAS and reinstall DSM — do not treat "Mode 2 reset" as a first step for general users.

QNAP (official FAQ, "My RAID is in degraded mode. What should I do?"): Degraded mode means one or more drives have failed; the official advice is to replace the failed drive as soon as possible. The failed drive should be hot-swapped without powering off: prepare a new hard drive at least as large as the failed drive; remove the failed drive; insert the new drive; check the RAID Group information in the Storage Pool or Static Volume management window — the status should show Rebuilding with progress; after the rebuild completes, the Status LED turns green and the volume status becomes Ready. If the rebuild does not start automatically, you can first set the new drive as a spare disk. Note: do not treat "hot-swap" as universal — hot-swap support varies by model, so confirm against your own model's manual.

<a id="sec-7"></a>

6. When to DIY, When to Call Professional Recovery

The dividing line is not "do you know computers" but the answers to four questions: How many drives failed — does it exceed the fault-tolerance limit of that RAID level? Are there signs of physical damage (abnormal noise, a drive not detected)? Have any operations already been performed that may have overwritten data (rebuild, initialise, chkdsk / fsck)? Is the data in the array business-critical, where further loss cannot be tolerated?

Relatively suitable for handling via the vendor's official process: only one drive has failed, the array shows degraded status but remains accessible, and a backup exists. Do not treat "a single-drive failure is always safe" as a conclusion — even with a single-drive failure, the array has no redundancy during the rebuild, and if a second drive fails during the rebuild, the risk rises.

Stop and consider professional recovery when:

• Multiple drives have failed at the same time, exceeding the RAID level's fault-tolerance limit;
• A drive has abnormal noises (clicking / grinding) or is not detected at all;
• A rebuild has failed / a previous recovery attempt has already made things worse;
• You are unsure of the RAID settings or the cause of failure and would be guessing parameters;
• The array configuration is non-standard / customised;
• The data is business-critical / legally relevant and cannot tolerate trial and error.

DIY vs Stop and Call a Professional (Comparison)

Distinguish logical from physical damage: when the main problem is logical in nature and most or all drives still spin up and are detected by the system, software-based handling is the more suitable end of the spectrum; abnormal noise, a drive completely undetected, or multiple simultaneous drive failures point towards a professional lab. Do not treat "logical damage = DIY-able" as a conclusion — there is a grey zone in the middle where you should stop and ask for professional advice.

<a id="sec-8"></a>

7. How Is Pricing Calculated? Why Individual Quotes?

RAID recovery pricing varies with drive count, failure type, capacity, whether cleanroom work is needed, and urgency — an inspection is required before quoting. This article gives no specific HKD figures, because research found no Hong Kong company publicly listing HKD RAID recovery prices on its website; an old 2014–2015 promotional price list has expired and cannot be cited as a current price.

Overseas prices are for reference only and cannot be taken as Hong Kong prices. For example, US-based $300 Data Recovery publicly lists per-drive RAID recovery pricing (US$300 per drive at 1TB or below, US$500 per drive over 1TB up to 6TB, US$700 per drive over 6TB, with every drive in the array charged). Dave's Computers in New Jersey lists RAID / NAS recovery from US$599 in its October 2026 price snapshot. These are prices published on those companies' own websites, in their original currencies — for reference only; Hong Kong pricing requires an individual quote.

<a id="sec-9"></a>

8. Prevention: So Next Time Is Less Frightening

• Complete 3-2-1 backup: 3 copies of the data (original + two backups), 2 different storage media, 1 offsite copy. RAID keeps the system running through a drive failure, but it copies deletions and corruption instantly; sync is not backup — sync propagates unwanted changes, and a business needs versioned restore points, a retention period, and independent restore capability. Test your backups by restoring — "never tested a restore" is one of the most dangerous mistakes, according to backup-practice write-ups. Backup systems should use separate credentials, not shared with everyday admin accounts. Do not assume 3-2-1 means you will never lose data — no solution can guarantee that.
• Regular RAID scrubbing: QNAP recommends RAID scrubbing at least once a month, scanning RAID 5 or RAID 6 group sectors and automatically attempting to repair detected errors; data scrubbing as a background process can detect and recover from UREs, effectively reducing the risk of a URE causing a double-drive failure during a rebuild.
• Monitoring + alerts: watch for early warning signs — sudden slowdowns, clicking sounds from a drive, critical S.M.A.R.T. alerts in the system log. Set up NAS notifications (e.g. email alerts) in normal times so drive warnings are seen early, instead of only discovering the problem when files will not open.
• UPS protection: keep stable power during rebuilds, ideally through a tested UPS.
• Proper records + replacement-drive planning: keep good records of the RAID configuration, a replacement-drive plan, and monitor abnormal drive behaviour. Store the RAID level, drive order, controller model and similar details somewhere other than the NAS itself (e.g. the owner's phone notes or paper), because when a failure happens, the files inside the NAS may not open at all.

Common Scenarios for Hong Kong SMEs (generic scenarios, no fictional cases)

• Accounting system on the NAS, NAS fails: follow the stop-writing + record steps above; do not hit rebuild or restart repeatedly in a panic. First establish whether there is a recent backup, where it is, and whether it can be restored on another computer and checked against the latest bookkeeping. Even after a successful restore, verify the accounting data's completeness — do not assume restored means complete.
• File server on the NAS, the whole company cannot open shared folders: the downtime impact is not just "cannot open files" — it includes being unable to reply to clients, issue quotations, or follow up orders. If a temporary workaround is needed, restore from backup to another storage device; do not "repair while using" the failed array directly.
• Beeping / warning from the NAS just before closing time: first read the warning — is it a single-drive warning or multiple drives warning at once? Degraded, or already inaccessible? Do not press a button with unclear consequences just because you "want to go home"; but do not skip recording and backing up out of fear of "breaking it". At minimum: photograph the warning screen, record which drive, confirm backup status, and keep the NAS in a safe state.
• No IT staff in the company — only the owner / clerk looks after the NAS: provide a checklist a non-technical person can follow — stop writing, take photos, copy down error messages, do not press rebuild / initialise, confirm backups, and provide the information from the steps above when contacting support. The point is knowing how to tell which situations can follow the vendor's official process and which require stopping.

<a id="sec-10"></a>

FAQ

Q1: When RAID fails, is the data gone immediately?

Not necessarily. With a single-drive failure and the array still in degraded status, data is usually still accessible; multi-drive failure, lost controller configuration, or destructive operations already performed make the situation much more serious. Stopping immediately and recording the current state is the key to preserving data.

Q2: Can I press rebuild myself?

Only on the precondition that it is confirmed that just one drive has failed, the remaining drives are healthy, and a backup exists — and then only by following the vendor's official process. After a first rebuild fails, do not attempt a second one; stop immediately and treat it as a professional-recovery situation.

Q3: One drive in RAID 1 failed — do I need to replace it immediately?

Yes. After one drive in RAID 1 fails, the array has no redundancy left, and the surviving drive is the only copy; if it fails during the rebuild, the whole array can collapse. Before swapping, confirm backup status; before pulling, check the bay number and serial number — do not pull the wrong drive.

Q4: How long does a RAID 5 rebuild take?

It can take many hours, sometimes more than a day for large drives, depending on drive capacity, array size, and system load. Until the rebuild completes, the array stays degraded and fragile; during the rebuild, keep power stable and avoid heavy workloads and unnecessary restarts.

Q5: No backup — can the data still be recovered?

Possibly, but do not try random fixes yourself. With single-drive failure, logical damage and similar situations, professional recovery may be able to rebuild the array on copies and extract files; multi-drive physical failure, repeated rebuilds, or initialisation make it much harder. Either way, the first step is always stop writing, record, and keep all drives.

Q6: Why does RAID recovery need an individual quote instead of a stated price?

Because pricing depends on drive count, failure type (logical or physical), capacity, whether cleanroom work is needed, and urgency — every job is different. Research found no publicly listed HKD RAID recovery price list in Hong Kong, so this article will not invent any HKD figures; the correct approach is to quote after inspection and start only after approval.

Q7: One drive in RAID 0 failed — what now?

RAID 0 has zero fault tolerance; one drive failure can disable the whole array. Stop immediately, do not perform any write operations on the array, and consider professional recovery; at the same time accept the reality that "without a backup, recovery may be impossible", and close the gap afterwards with 3-2-1 backup.

<a id="sec-11"></a>