Damaged HDD: 5 mistakes that reduce recovery chances

Damaged HDD on a workbench in a Warsaw data recovery case

A drive rarely fails at a convenient moment. It may start clicking on a Sunday evening in an apartment in Mokotów, freeze a laptop during a copy job in Wola, or appear as RAW just before a company deadline. In the first hour, the goal is not to “try everything”. The goal is to avoid changing the evidence that still exists on the platters.

These five mistakes are the ones we most often have to untangle in the laboratory. Some are Windows habits, some are internet myths, and some are just stress reactions. The safer option is usually quieter: stop writes, note the symptoms and decide whether the case is logical, mechanical or already a lab job.

Mistake 1: running CHKDSK or automatic repair on an unstable drive

CHKDSK can repair file-system structures on a healthy disk. On an HDD with bad sectors, slow reads or disconnects, it may rewrite the MFT, move entries to found folders or leave a half-repaired file system behind. That can make later reconstruction harder, especially when the drive is already losing sectors.

If Windows says the drive needs repair, first ask what the hardware is doing. Does copying stop with a CRC error? Does the disk disappear and come back? Does the computer hang when the drive is attached? In those cases, read the dedicated warning about not using CHKDSK on a damaged drive before letting the system change the original media.

Mistake 2: installing recovery software on the same disk

Recovery software is not the problem by itself. The problem is writing to the same storage device that contains the missing files. The installer, browser download, temporary files and output folder can overwrite the very sectors you want to recover.

On an SSD or NVMe drive, the risk is higher because TRIM and background cleanup may remove deleted-file references quickly. If the device is stable and you still choose a software attempt, run the tool from another disk and save results elsewhere. When the drive is noisy, RAW, disconnecting or reporting I/O errors, move to HDD data recovery or SSD/NVMe data recovery instead of experimenting on the source.

Mistake 3: formatting or initialising the disk so it appears again

RAW, “unknown” and “not initialised” messages are system interpretations, not proof that the files are gone. Formatting, creating a new partition or initialising MBR/GPT writes fresh structures to the device. Sometimes recovery is still possible afterwards, but the case becomes less clean.

If Windows asks to format the drive, do not confirm just to “unlock” it. The same applies to Disk Management prompts asking for initialisation. Compare the symptom with our notes on the drive needs to be formatted message, RAW drives and unknown or uninitialised disks.

Mistake 4: powering on a drive with mechanical symptoms again and again

Clicking, repeated spin-up attempts, squealing, scraping, BIOS freezes and a drive that vanishes after warming up are mechanical warning signs. More power cycles can turn a recoverable head or surface problem into wider platter damage.

Do not open the HDD at home, do not put it in a freezer, and do not keep copying “until it finally catches”. If the drive clicks on startup, keep it powered off and describe the exact sound, device model and last successful access. The guide on a drive clicking on startup explains when tests should stop.

Mistake 5: trusting a backup before checking what is inside

“We have a backup” is only useful if it contains the current files and opens correctly. We see cases where the backup was months old, synced the wrong folders, or lived on the same failing NAS as the original data.

Before formatting, rebuilding a RAID array or running repair tools, verify the backup on another device. If the array or server holds the only usable copy, treat it as a business incident and review RAID/NAS data recovery before starting automatic rebuilds.

What to write down before you call a laboratory

  1. Exact symptoms: sound, messages, capacity shown, speed drops, disconnects.
  2. What happened before the failure: fall, power cut, update, rebuild, liquid, format, deleted partition.
  3. Device details: HDD model, enclosure, laptop model, RAID level or NAS model.
  4. Actions already taken: CHKDSK, format prompt, software scan, disk swap, rebuild attempt.
  5. Most important folders: accounting, photos, client data, virtual machines, project files.

A better first move

If several symptoms appear together, stop the chain of attempts. Prepare a short case description, use the initial assessment form or call the Warsaw laboratory before more sectors, logs or file-system structures are overwritten.

Choose the recovery procedure by device type

If the drive is unstable, noisy, shown as RAW or asking for formatting, match the next step to the storage type before making another write attempt.

Not sure whether another attempt is safe?

Tell us what the drive shows, what it sounds like and what has already been tried. We will help you choose the safest next step before the evidence changes.

Call the lab