lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date:   Tue, 10 Oct 2017 16:43:58 -0600
From:   Andreas Dilger <adilger@...ger.ca>
To:     Drake Cai <dcai.xff@...il.com>
Cc:     linux-ext4@...r.kernel.org
Subject: Re: Odd EXT4 corruption

On Oct 3, 2017, at 12:10 AM, Drake Cai <dcai.xff@...il.com> wrote:
> 
> As I was trying to install a new distribution I accidentally gdisk'd
> /dev/sda instead of /dev/sdc and deleted all of the partitions off of
> my device. I also wrote to the device a new GPT partition table label
> with a 2G EFI partition and a Linux Filesystem partition that filled
> the rest of the disk. I later then deleted the two partitions and made
> an ext4 partition that filled the rest of the disk. At this point
> forwards, I decided to do a complete dd backup of the 1.8TiB drive and
> am now working on the backed up identical drive (also 1.8TiB). I never
> mkfs'ed the device so the data in theory should all be there.
> 
> I have also used TestDisk and PhotoRec respectively. TestDisk fails to
> detect this weird situation and PhotoRec generates way too many files
> for it to be of use.
> 
> The disk in question was originally formatted with Gnome's Disk
> Utility with no partitioning and just an ext4 filesystem. Mountings of
> the device on the superblock backups blocks 65536, 131072, etc (but
> not 32768) are successful.
> 
> However, after mounting, while the first directory of folders exists,
> when any following directories are entered ls shows a "Structure needs
> cleaning" error for the majority of the files and folders. Running
> e2fsck with the superblock backups failed to fix the problem but left
> me with 29GiB of the original 1TiB of data (the folders & files
> without "Structure needs cleaning" errors) of the 1.8TiB drive.
> Lost+Found did not contain any of the remaining data either.
> 
> A friend on IRC tried reproducing my exact steps (creating drive,
> adding content, gdisking and attempting to mount/fsck) on his own
> setup and also came to the same error of "Structure needs cleaning"
> for directories.
> 
> I did a little more digging into GPT and EXT4 and realized using gdisk
> meant I overwrote the first 33792 bytes of the EXT4 filesystem. I'm
> having a little trouble understanding whether or not I overwote the
> GDT blocks which I was not using or if I overwrote some crucial data &
> inode bitmaps or possibly even inode table blocks. I should have only
> overwritten 8 or so blocks to my judgement. I'm not too sure whether
> or not GDT blocks come before bitmap or inode table blocks or are
> variable to be more precise.

If you are positive you have rebuilt the partition table correctly, then
it is possible that you corrupted the ext4 journal and this is causing
the error message to be hit.

> I've attached below the output of tune2fs of the device.

It would be useful to get the actual e2fsck error message as well.

> So far this is as far as I have got. Does anyone have any ideas on how
> I could salvage the rest of the data? Any help is appreciated.

Since you already have a copy of the filesystem, I would suggest to
try removing the ext4 journal ("tune2fs -O ^has_journal /dev/XXX" should
be OK) and then run "e2fsck -fy" to repair the filesystem.

Failing that, you could try stracing the e2fsck run to see where the
"EUCLEAN" error ("Structure needs cleaning") is generated, and why.
This error doesn't appear to be in the e2fsprogs source code.

Cheers, Andreas

> Thank you.
> 
> ====================
> 
> tune2fs 1.43.4 (31-Jan-2017)
> Filesystem volume name:   <none>
> Last mounted on:          <not available>
> Filesystem UUID:          57e8dc69-003e-4d43-a912-adcbc09d8110
> Filesystem magic number:  0xEF53
> Filesystem revision #:    1 (dynamic)
> Filesystem features:      has_journal ext_attr resize_inode dir_index
> filetype extent 64bit flex_bg sparse_super large_file huge_file
> dir_nlink extra_isize metadata_csum
> Filesystem flags:         signed_directory_hash
> Default mount options:    user_xattr acl
> Filesystem state:         clean
> Errors behavior:          Continue
> Filesystem OS type:       Linux
> Inode count:              122101760
> Block count:              488378646
> Reserved block count:     24418932
> Free blocks:              480431423
> Free inodes:              122101749
> First block:              0
> Block size:               4096
> Fragment size:            4096
> Group descriptor size:    64
> Reserved GDT blocks:      1024
> Blocks per group:         32768
> Fragments per group:      32768
> Inodes per group:         8192
> Inode blocks per group:   512
> Flex block group size:    16
> Filesystem created:       Mon Oct  2 20:55:33 2017
> Last mount time:          Mon Oct  2 20:56:09 2017
> Last write time:          Mon Oct  2 20:56:16 2017
> Mount count:              2
> Maximum mount count:      -1
> Last checked:             Mon Oct  2 20:55:33 2017
> Check interval:           0 (<none>)
> Lifetime writes:          1054 MB
> Reserved blocks uid:      0 (user root)
> Reserved blocks gid:      0 (group root)
> First inode:              11
> Inode size:              256
> Required extra isize:     32
> Desired extra isize:      32
> Journal inode:            8
> Default directory hash:   half_md4
> Directory Hash Seed:      cbf62753-15ca-4878-9a9e-6594eeaabc85
> Journal backup:           inode blocks
> Checksum type:            crc32c
> Checksum:                 0x87d53c7b


Cheers, Andreas






Download attachment "signature.asc" of type "application/pgp-signature" (196 bytes)

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ