Nothing went wrong with dd itself — it copied exactly what it was pointed at: raw blocks. A block-level clone cannot turn an NTFS partition into an ext4 one, and when you point dd at a freshly formatted ext4 partition, the first thing it overwrites is that partition's own filesystem metadata. The result is a target your file manager cannot present as files. To move your data to ext4, mount both filesystems and copy the files with rsync.
What dd really does — and why the target looked wrong
dd knows nothing about files, folders, or filesystems. It reads a device sector by sector — boot sectors, journals, allocation tables, free space, everything — and writes those bytes to the destination. When you run a command like dd if=/dev/sdc5 of=/dev/sdd1, the destination's ext4 superblock is destroyed within the first few kilobytes and replaced with the source partition's raw contents.
That explains every oddity you observed:
- The target no longer contains a valid ext4 filesystem, so the file manager shows an empty folder, a mount error, or nonsense depending on what the kernel still has cached.
- Even a "successful" clone would be readable only as the source's filesystem type — you would have simply made a second NTFS partition, which defeats the purpose.
- A clone carries the source's filesystem UUID, so with both drives attached, the system can mount the wrong one at boot.
- Usage figures won't match your file totals, because
ddcopies allocated blocks and slack space, not a file tree. - Cloning a smaller partition onto a larger one strands the leftover space until you resize; cloning onto a smaller one truncates data.
Copy files, not blocks, when changing filesystems
A filesystem conversion means reading each file through the NTFS driver and writing it through the ext4 driver. That is inherently a file-level operation. Here is a safe sequence:
- Identify the devices — never trust remembered letters. USB device names shift between boots and ports. Run
lsblk -fand confirm which device is the Seagate and which is the target by size, label, and filesystem. - Back up anything irreplaceable first. The next steps destroy the target, and a mistake with device names can destroy the source.
- Format the target while it is unmounted:
If you get a "target is busy" error, close any file-manager windows showing that drive and retry.sudo umount /dev/sdX1 sudo mkfs.ext4 /dev/sdX1 - Mount both filesystems. Desktop environments usually auto-mount on plug-in; otherwise:
sudo mkdir -p /mnt/source /mnt/target sudo mount /dev/sdY1 /mnt/source sudo mount /dev/sdX1 /mnt/target - Copy with rsync:
The flags mean:sudo rsync -rahAEXHP --info=progress2 /mnt/source/ /mnt/target/-rrecursive,-aarchive mode (preserves structure and timestamps),-Aand-Xkeep ACLs and extended attributes,-Hpreserves hard links,-Ekeeps executability, and-h/-Pgive readable output with progress and resume. Note the trailing slash on the source: with it, rsync copies the contents into the target; without it, you would create a nested/mnt/target/sourcefolder.
One expectation to set: NTFS stores no Linux ownership or permission bits, so files arrive with defaults and will be owned by root when run under sudo. Run chown/chmod afterwards if that matters to you.
Verify before you touch the original drive
- Compare totals:
du -sh /mnt/source /mnt/target - Compare file counts:
find /mnt/source -type f | wc -lagainst the same command on the target. - Re-run the same rsync command. It scans both trees and transfers only differences, so a second run that finishes almost instantly is strong evidence the copy is complete.
- Spot-check by opening a few large files directly from the target.
Only after verification should you reformat or repurpose the old NTFS drive — or better, keep it as a backup for a while.
Is ext4 the right destination?
If the drive is used exclusively with Linux, ext4 is a sound choice: it is far more at home on Ubuntu than NTFS, and Linux's ntfsfix can only clear a dirty flag rather than actually repair the filesystem — which is why every hiccup sent you to a Windows machine for chkdsk. The trade-off is that Windows cannot read ext4 without third-party drivers. If the drive might ever be plugged into a Windows PC, consider exFAT (universally readable, but no journaling or permissions) or staying with NTFS. If you commit to ext4, add the drive to /etc/fstab using the new UUID shown by lsblk -f so it mounts reliably.
Preventing the next dd accident
- Treat
ddandmkfsas commands that destroy their target. Runlsblk -fimmediately beforehand and re-read theof=argument before pressing Enter. - Use file-level tools (
rsync,cp -a, or a dedicated backup program) for migrations and routine backups. - Reserve
ddfor what it is good at: writing images, wiping devices, and cloning a disk to an identical one. If you do clone, give the copy a fresh identity afterwards — for ext4,sudo tune2fs -U random /dev/sdX1— so the two can never be confused at mount time. - For full-disk imaging, consider Clonezilla, which understands filesystems and copies only used blocks.
Frequently Asked Questions
Why did my cloned drive show strange or empty contents?
Because the clone overwrote the ext4 metadata that made the target readable, replacing it with raw source bytes. The partition is only interpretable as the source's filesystem type, and the file manager's cached view no longer matches the disk.
Can dd ever convert NTFS to ext4?
No. dd never interprets filesystems, so it can only produce an identical copy — same type, same UUID, same layout. Conversion requires reading files through one driver and writing them through another, which is what rsync does.
Does rsync preserve everything from an NTFS drive?
It preserves the file tree and timestamps. NTFS has no native Linux ownership or mode bits, so those are assigned defaults; ACLs and extended attributes are preserved only where they exist.
What is the difference between /mnt/source and /mnt/source/ in rsync?
A trailing slash means "copy the contents of this directory." Without it, rsync copies the directory itself, creating an extra nesting level on the target.
Both drives have the same UUID after a dd clone — how do I fix that?
Assign a new one: sudo tune2fs -U random /dev/sdX1 for ext4. Or avoid the problem entirely by using a file-level copy instead of a block clone.
Is dd the wrong tool for backups in general?
Not wrong, just specialized. It excels at writing images and producing exact replicas of identical hardware, but for ongoing backups and filesystem migrations, file-level tools are faster, safer, and far easier to verify.