How to recover data from a crashed hard drive for free

What you need

Gather these before starting

Step-by-step recovery

Six steps. The whole process takes 1-4 hours depending on drive size and damage severity.

1. 1

   Create a SystemRescue USB

   Download SystemRescue and flash the ISO to a USB drive (2 GB minimum). On Windows, use Rufus. On Mac, use balenaEtcher. On Linux, use dd. SystemRescue includes ddrescue out of the box; no additional packages to install.

   Verify your download checksum against the one listed on the SystemRescue website before flashing. A corrupted boot USB will waste time.
2. 2

   Boot into SystemRescue

   Plug in the SystemRescue USB and reboot. Press F12 (or F11) on PCs to reach the boot menu. On Macs, hold Option and select the USB stick. Choose the default SystemRescue environment when prompted. You will land at a Linux command line.

   If your machine does not show the USB in the boot menu, check that Secure Boot is disabled in BIOS/UEFI settings.
3. 3

   Identify source and target drives

   Linux names drives differently than Windows. Run one of these commands to list every connected drive with its size:

   # List drives with sizes
   lsblk
   # Or show partition tables
   fdisk -l

   Copy
   Triple-check drive letters before proceeding. Writing to the wrong drive destroys its contents. Match by size and partition layout, not by name alone.
4. 4

   First pass: grab the easy data

   This copies every sector that reads within a reasonable time and skips anything slow or unresponsive. Replace /dev/sdX with your source and /dev/sdY with your target.

   ddrescue -f -n -a 5120000 /dev/sdX /dev/sdY logfile.log

   Copy
   -nNo scraping (fast pass only)

   -a 5120000Skip regions below ~5 MB/s

   logfile.logTracks progress for resuming
5. 5

   Second pass: retry the difficult sectors

   Warning: Repeatedly retrying bad sectors on a mechanically degrading drive generates extreme thermal and mechanical stress on the read-head assembly that can cause permanent head collapse and total data loss. The following command should only be used if the drive is NOT clicking, grinding, or dropping offline during the first pass.

   Once the fast pass finishes, run this command to go back for the sectors that were skipped. The logfile tells ddrescue exactly what still needs reading.

   ddrescue -f -d -r1 /dev/sdX /dev/sdY logfile.log

   Copy
   -dDirect disk access (bypasses cache)

   -r1Retry bad sectors only once to minimize mechanical stress

   logfile.logContinues where pass 1 left off
   Stop here if needed: If you encountered a high number of bad sectors in the first pass, stop here and consult a professional lab. Continuing to retry on a drive with extensive bad sectors risks total head failure and permanent data loss.
6. 6

   Mount the clone and copy your files

   Disconnect the failing source drive. Mount the target (your clone) and copy files to a third healthy drive. If the file system is corrupted, run fsck or chkdsk on the clone only.

   Mount read-only first: mount -o ro /dev/sdY1 /mnt/recovery prevents accidental writes to the clone while you inspect the file system.

Common mistakes that destroy data

What professional recovery looks like

When DIY cloning is not enough, the next step is firmware-level access and physical head replacement. Here is a Seagate Rosewood case from our Austin lab.