[TriLUG] Trouble with NTFS partition on Linux

Paul Boyle pboyle at uwo.ca
Tue Feb 11 13:46:21 EST 2014


Hi,

I have an hard drive for archiving data.  Rather than
reformatting as an ext4 file system, I just went with the hard drive'0s
NTFS file system. So far, this hadn't caused any problems.  Recently I
had to copy a bunch of files from an old Linux system (Redhat 8-ish
don't ask, it can't upgraded) over NFS onto one of these hard drives
which is attached to a newer Linux box (running OpenSuSE 12.2).
Eventually, I deleted the bunch of files which came from the NFS copy
from the external hard drive except for one recalcitrant directory
which can't be deleted.

The external hard drive mounts and unmounts without difficulty, and in
all other respects the drive seems fine.

Here is some output regarding the problem directory (named n13052).
Annotations are in /* C style comments */

xray:/ragogna_nonius_data% ls -la --color=no
total 472
drwxrwxrwx 1 root root   4096 Feb 10 12:47 .
drwxrwxrwx 1 root root   4096 Jan  6 15:28 ..
drwxrwxrwx 1 root root 475136 Jan 10 16:06 n13052

/* notice the anomalous block size of 475136 for the directory */

/* below some conventional file operations, all of which fail */

bravais:/home/backup/Primary/ragogna_nonius_data # ls -l n13052/
ls: reading directory n13052/: Input/output error
total 0
bravais:/home/backup/Primary/ragogna_nonius_data # rmdir n13052/
rmdir: failed to remove ‘n13052/’: Directory not empty
bravais:/home/backup/Primary/ragogna_nonius_data # rm -rf n13052/
rm: cannot remove ‘n13052’: Directory not empty
bravais:/home/backup/Primary/ragogna_nonius_data #

/* using rsync also shows errors (below): */
 rsync -av Primary/ Mirror/
sending incremental file list
rsync: readdir("/home/backup/Primary/ragogna_nonius_data/n13052"):
Input/output error (5)

sent 6222 bytes  received 26 bytes  12496.00 bytes/sec
total size is 145898607714  speedup is 23351249.63
rsync error: some files/attrs were not transferred (see previous
errors) (code 23) at main.c(1052) [sender=3.0.9]

Doing these commands, I get the following message
in /var/log/messages:

Feb 11 13:15:52 bravais ntfs-3g[31069]: ntfs_mst_post_read_fixup_warn:
magic: 0x08458b10  size: 4096   usa_ofs: 25997  usa_count: 23539: Invalid argument
Feb 11 13:15:52 bravais ntfs-3g[31069]: Actual VCN (0xd845c7bc558bec)
of index buffer is different from expected VCN (0x1c) in inode 0xc26e.
Feb 11 13:16:30 bravais ntfs-3g[31069]: ntfs_mst_post_read_fixup_warn:
magic: 0x08458b10  size: 4096   usa_ofs: 25997  usa_count: 23539: Invalid argument
Feb 11 13:16:30 bravais ntfs-3g[31069]: Actual VCN (0xd845c7bc558bec)
of index buffer is different from expected VCN (0x1c) in inode 0xc26e.
Feb 11 13:30:23 bravais ntfs-3g[31069]: ntfs_mst_post_read_fixup_warn:
magic: 0x08458b10  size: 4096   usa_ofs: 25997  usa_count: 23539: Invalid argument
Feb 11 13:30:23 bravais ntfs-3g[31069]: Actual VCN (0xd845c7bc558bec)
of index buffer is different from expected VCN (0x1c) in inode 0xc26e.


I've unmounted the disk and used the 'badblocks' command.  There are no
badblocks, at least according to that program.

I've looked at the ntfsprogs(8) man page, but nothing seems to address
this particular problem.

I've googled around for these errors, but haven't found too much
helpful information.

I would appreciate any pointers which can help me get rid of this
directory.

Thanks,

Paul




-- 
Paul D. Boyle, Ph. D.
Manager, X-ray Facility
Department of Chemistry
Western University
London, ON N6A 5B7
Canada


More information about the TriLUG mailing list