How do I clone a failing drive with ddrescue before attempting any recovery?

A failing hard drive rarely gives you a second chance. Clicking sounds, slow reads, and disappearing folders are signs the disk weakens every time you power it on. Running recovery software directly on it may push it over the edge.

The safest first step is always the same: make a sector-by-sector clone with GNU ddrescue, then do all recovery work on the clone.

Quick Answer

Boot from a Linux live USB, connect a healthy destination drive that is at least as large as the failing drive, and run ddrescue with a map file so it copies all the easy, healthy sectors first. A typical first pass is sudo ddrescue -f -n /dev/sdX /dev/sdY rescue.log, where sdX is the failing source and sdY is the destination. Never work on the original drive again after cloning — run all file recovery tools on the clone only.

Why Cloning Must Come Before Any Recovery

Most recovery programs scan a drive again and again, retry bad areas, and keep the drive busy for hours. A failing drive cannot handle that. Every extra read can heat up weak heads, spread platter damage, or trigger more bad sectors.

ddrescue works differently. It grabs the easy data first, skips damaged areas for later, and records what it has copied. If the drive dies halfway, you can resume from the same point. Once cloning finishes, switch the original off and leave it alone.

Clone, Image File, and Copy Are Not the Same Thing

Copying files in Windows Explorer only copies files the system can already see. It misses deleted files, damaged partitions, and empty-looking space that still holds recoverable data. A clone or a full disk image copies every sector, including free space, partition tables, and file system structures. That complete copy is what recovery software needs to do its best work.

What You Need Before You Start

• A destination drive that is the same size or larger than the failing drive, and fully healthy.
• A Linux live USB, such as Ubuntu, or a rescue system that includes GNU ddrescue.
• Enough time and stable power. Use a direct SATA or good quality USB dock connection, not a loose cable.
• A USB stick or separate drive to keep the ddrescue map file safe if you are imaging to a file.
• Patience. A badly damaged drive can take many hours or even days to clone.

Warning: ddrescue will overwrite everything on the destination drive. There is no undo. Back up anything important from the destination first, and label both drives with tape if it helps you avoid confusion. Double-check the source and destination order before you press Enter. Reversing them will overwrite the failing drive with the empty destination, and that data loss cannot be fixed.

Step-by-Step: Clone a Failing Drive With ddrescue

Step 1: Stop Using the Failing Drive

1. Shut the computer down cleanly if you can. Do not run check disk, defragment, or antivirus scans.
2. Disconnect the failing drive until you are ready to clone. Do not keep rebooting into Windows on that drive.
3. If the drive makes loud clicking, grinding, or beeping sounds and does not appear in the BIOS at all, stop and consider a professional lab. Software cannot fix a physically seized motor or broken heads.

Step 2: Boot Into Linux and Identify Both Drives

1. Boot from your Linux live USB and choose the Try option, not Install.
2. Connect the failing drive and the destination drive.
3. Open a terminal and run lsblk -o NAME,SIZE,MODEL,SERIAL and sudo fdisk -l to list all drives.
4. Identify each drive by its size, model, and serial number — never by letter alone, because Linux device names can change between boots.
5. Write down which device is the source, for example /dev/sda, and which is the destination, for example /dev/sdb.

Step 3: Install ddrescue if Needed

Many rescue systems include it already. On Ubuntu live, you may need to enable the universe repository and run sudo apt update followed by sudo apt install gddrescue. Note the package name on Debian and Ubuntu is gddrescue, but the program itself is still called ddrescue. You can confirm correct usage any time in the official GNU ddrescue manual and documentation.

Step 4: Run the First Fast Pass

1. Make sure no partitions from either drive are mounted. Run mount | grep sd to check, and unmount if needed.
2. Run the first pass with no scraping and no retries: sudo ddrescue -f -n /dev/sdX /dev/sdY rescue.log
3. Replace sdX with your failing source and sdY with your destination. Keep rescue.log on persistent storage, not in a folder that disappears when you reboot the live system.
4. Watch the output. ddrescue shows how much has been copied, the current read speed, and the size of error areas.
5. Let this pass finish even if it takes hours. It is designed to save the healthiest data first, while the drive is at its strongest.

The -f option allows ddrescue to write directly to a device. The -n option skips the scraping phase, so ddrescue does not waste early time on the hardest spots. The map file, rescue.log, is essential. Without it, ddrescue cannot resume intelligently.

Step 5: Run a Second Pass to Retry Damaged Areas

1. After the first pass completes, run: sudo ddrescue -f -r3 /dev/sdX /dev/sdY rescue.log
2. The -r3 option retries each bad area up to three times. You can increase the number if the data is very valuable, but more retries also stress the drive more.
3. If the drive gets hot, starts clicking more, or disappears, stop, let it cool, and continue later with the same map file.
4. Repeat only until extra runs stop gaining meaningful new data. At some point, the remaining sectors are simply unreadable by software.

Step 6: Disconnect the Original and Work Only on the Clone

1. Shut down and physically disconnect the failing drive. Label it and store it safely.
2. Boot normally or attach only the clone, and run your recovery software against the clone.
3. If you cloned to an image file instead, mount or open that image with your recovery tool and scan the image, never the original.

Our companion guide on recovering data from a drive that Windows no longer detects explains what to do next once you have a safe clone to scan.

Useful ddrescue Options Worth Knowing

• --reverse — reads the drive backwards, which sometimes helps when damage is at the start of the disk or the heads struggle in one direction.
• --timeout=30s — prevents ddrescue from waiting forever on a single stuck sector.

Do not add many options at once if you are a beginner. A clean first pass with -f -n and a proper map file solves most cases.

Common Mistakes That Destroy Recoverable Data

• Reversing source and destination: This is the worst mistake. Always verify by serial number and size, and read the command aloud before running it.
• Forgetting the map file: Without rescue.log, a power cut or crash forces you to start again and stresses the drive for no reason.
• Running recovery software on the original: Scans, previews, and test recoveries all belong on the clone. The original should stay powered off.
• Using a destination that is too small: Even if your files are only 200 GB, a 1 TB source needs a 1 TB or larger destination for a full sector clone.
• Freezing, hitting, or opening the drive: Old internet tricks cause more harm than good. Opening a drive outside a clean room almost always ends recovery chances.
• Giving up after the first pass: The first pass is meant to be incomplete. Its job is to save easy data fast. The retry passes are where difficult sectors are won back.

How to Know If Your Clone Is Good Enough

ddrescue reports the percentage rescued and the size of the error area at the end of each run. A rescue rate above 99 percent is excellent, and most files will usually open normally. Between 95 and 99 percent, many files will still be usable, but some photos, videos, or databases may be damaged where bad sectors fell inside them. Below that, recovery becomes partial, and you should set expectations accordingly. Always try opening a sample of important files from the clone before assuming success.

Keep the clone and map file until all recovered files are verified and backed up elsewhere.

Conclusion

Cloning a failing drive with ddrescue is not the glamorous part of data recovery, but it is the part that decides whether recovery succeeds at all. Boot from a live USB, identify your drives with total care, run a fast first pass with a map file, then retry the damaged areas gently. After that, put the original drive away and do every scan and repair on the clone only. That single habit — clone first, recover second — protects more family photos, business documents, and irreplaceable files than any recovery program ever will on its own.

Frequently Asked Questions

Can I run ddrescue from Windows instead of Linux?

ddrescue is a Linux tool, so boot from a Linux live USB without installing anything. The live session does not change your Windows drives and gives direct low-level access to the failing disk.

How long should a ddrescue clone take?

A healthy drive may clone in a few hours. A failing drive can take a day or longer because ddrescue slows down around damaged areas. If the rescued percentage is still climbing, patience is better than stopping early.

What if ddrescue shows 0 bytes rescued or the drive disappears?

Check cables, power, and a different port first, then confirm the drive appears in lsblk. If it is not detected at all or vanishes repeatedly, software recovery has likely reached its limit and a professional lab is safer.

Do I need the map file if the clone finishes in one sitting?

Yes, always use it. You cannot predict a power cut or crash. The map file costs nothing and avoids rescanning a weak drive from the beginning.

Can I use the destination drive for normal storage after cloning?

Not until recovery is finished and verified. The clone is your only safe working copy until files are recovered, tested, and backed up elsewhere.