31

Is there a way to fsck/chkdsk an NTFS drive from Linux?

Improve this question
18

Yes. This was handled by fsck on some releases. If the partition is not listed in /etc/fstab, then you will likely need to tell it the partition type. I've used this from a Linux CD to recover a partition Windows wouldn't boot from.

Based on the comments below, the actual fixing is done by the ntfsfix program. It should be available, even if there is no program to run a fsck on and NTFS file system.

Improve this answer
  • 8
    Although as noted in the man page for ntfsfix/fsck.ntfs, it is not a Linux version of chkdsk and only checks for certain kinds of obvious problems on the disk. – SabreWolfy Dec 19 '11 at 10:03
  • 4
    I assume from the above comment that fsck.ntfs was at some point handled by ntfsfix. Comments on a lower post indicate this may not be workable now as their arguments are not compatible. Anyway, in Debian 8.2, fsck.ntfs does not exist - but ntfsfix does and was able to fix a corrupted MFT I encountered. Thanks @SabreWolfy for the hint. – underscore_d Nov 5 '15 at 23:24
11

Unfortunately the ntfsfix tool is very limited compared to Microsoft's chkdsk. Try to get a Windows install going - preferably with the newest version of Windows as Microsoft is presumably constantly improving chkdsk (I hope..) New versions of Windows are often available for free as trials. If the problem is in a USB disk you can try installing Windows in something like VirtualBox and give the VM control of the USB device.

Improve this answer
  • 1
    Or use Ultimate Boot CD for Windows. It's a live Windows XP CD with lots of recovery tools. That's how I solved my problem (corrupted NTFS partition). I tried using the ntfsck program included in the Parted Magic live USB, but I got an error saying Boot sector: bad NTFS magic. Segmentation fault. – dario_ramos Apr 30 '12 at 23:23
  • 1
    maybe limited for some people, but it fixed my corrupted MFT just fine :-) [until the next time, perhaps! i'm not sure about this drive...] – underscore_d Nov 5 '15 at 23:22
  • UBCD has many qualities but I should not describe it as a Windows XP CD. – Graham Perrin Jul 16 '16 at 14:33
  • 1
    Hyper-V Server is a completely free (beer) but heavily cut-down version of Windows that's designed for running VMs. The installer should contain all the usual utilities like chkdsk/diskpart, etc. – mwfearnley Dec 4 '19 at 13:22
  • 1
    Take care using Windows 'chkdsk /f' on an NTFS used by Linux. I just followed this advice to fix errors on my external 2TB backup drive. It deleted the names of 268 files and directories (incl. all my incr. backups) due to invalid characters (from NTFS viewpoint - most often ':' used in timestamps). I reverted it all using the log file, but it took half a day of coding. Some but not all Linux system-calls block writing such invalid names on an NTFS. I use rsync for backups, which silently copies such invalid filenames - normally useful, until you need to fix other problems on the disk... – Bob Briscoe Jun 15 '20 at 11:30
4

If you have NTFSProgs installed, you should be able to run fsck.ntfs or fsck -t ntfs to fsck an NTFS drive.

Improve this answer
  • 2
    Running Debian 6.0.3, just installed ntfsprogs, no fsck.ntfs installed, and when running fsck with -t ntfs, get "fsck.ntfs not found". So don't think the two are related. – Zayne S Halsall Dec 17 '11 at 15:00
  • 4
    @ZayneSHalsall In Debian, it happens to be a symlink to ntfsfix which is in ntfsprogs. I guarantee you that you will be able to run ntfsfix if ntfsprogs installed properly. – squircle Dec 27 '11 at 5:51
  • 4
    Newer versions of Debian and Ubuntu don't have this symlink. You can make it yourself, but the problem is that ntfsfix has a completely different set of arguments to the standard fsck ones. It's better to make fsck.ntfs (and fsck.ntfs-3) a script containing exec ntfsfix ${@:$#} (with a shebang of #!/bin/bash). – Neil Mayhew Mar 19 '15 at 22:31
  • @squircle thank you man sudo ntfsfix /dev/sda1 fixed my issue. fsck.ntfs -f /dev/sda1 and fsck -t ntfs -f /dev/sda1 exited early with fsck from util-linux 2.31.1 – Edward Oct 30 '20 at 13:22
2

ln -s /bin/ntfsfix /sbin/fsck.ntfs

fsck /dev/sdg1
fsck from util-linux 2.21.2
Mounting volume... OK
Processing of $MFT and $MFTMirr completed successfully.
Checking the alternate boot sector... FIXED
NTFS volume version is 3.1.
NTFS partition /dev/sdg1 was processed successfully.

On fedora 22, there is another binary :

lrwxrwxrwx. 1 root root 13 May 22 22:13 /usr/sbin/fsck.ntfs -> ../bin/ntfsck
Improve this answer
1

In my case, none of fsck & ntfsfix could fix the issue of my external NTFS mobile disk.

I end up to boot into windows os, and use follow steps to fix the disk issue.

Steps:

  • Check the identifier of the bad disk. e.g g
  • Open the "cmd" terminal of windows.
  • Input chkdsk <disk>: /f to fix it.
    Where <disk> is the identifier, e.g chkdsk d: /f
  • Then it would be fixed within seconds (For a disk of size 1Tb).

After years, finally I found something that windows can do easily, but linux can't ;)

@Update - vm solution

You can also create a windows vm (e.g win 7), then mount the mobile disk to windows vm, then you can also use chkdsk to fix the disk, it's similar as boot into a windows machine.

Improve this answer
  • 1
    I'd recommend using flag /x (unmount) for enforcing preventing access to the drive. Help manual says using /x implies also /f, so it helps for doing fixes the same way Linux does (fsck requires drives being unmounted). – SonicARG Jun 24 '18 at 1:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.