How can I recover an ext4 filesystem corrupted after a fsck?

Question

I have an ext4 filesystem on luks over software raid5. The filesystem was operating "just fine" for several years when I was beginning to run out of space. I had a 9T volume on 6x2T drives. I began upgrading to 3T drives by doing the mdadm fail, remove, add, rebuild, repeat process until I had a larger array. I then grew the luks container, and then when I unmounted and tried to resize2fs I was given the message the filesystem was dirty and needed e2fsck.

Without thinking I just did e2fsck -y /dev/mapper/candybox and it began spewing all kinds of inode being removed type messages (can't remember exactly) I killed e2fsck and tried to remount the filesystem to backup data I was concerned about. When trying to mount at this point I get:

# mount /dev/mapper/candybox /candybox
mount: wrong fs type, bad option, bad superblock on /dev/mapper/candybox,
       missing codepage or helper program, or other error
       In some cases useful info is found in syslog - try
       dmesg | tail  or so

Looking back at my older logs I noticed the filesystem was giving this error each time the machine booted:

kernel: [79137.275531] EXT4-fs (dm-2): warning: mounting fs with errors, running e2fsck is recommended

So shame on me for not paying attention :(

I then tried to mount using every backup superblock (one after another) and each attempt left this in my log:

EXT4-fs (dm-2): ext4_check_descriptors: Checksum for group 0 failed (26534!=65440)
EXT4-fs (dm-2): ext4_check_descriptors: Checksum for group 1 failed (38021!=36729)
EXT4-fs (dm-2): ext4_check_descriptors: Checksum for group 2 failed (18336!=39845)
...
EXT4-fs (dm-2): ext4_check_descriptors: Checksum for group 11911 failed (28743!=44098)
BUG: soft lockup - CPU#0 stuck for 23s! [mount:2939]

Attempts to restart e2fsck results in:

# e2fsck /dev/mapper/candybox 
e2fsck 1.41.14 (22-Dec-2010)
e2fsck: Group descriptors look bad... trying backup blocks...
candy: recovering journal
e2fsck: unable to set superblock flags on candy

At this point, I decided it best to order some more drives and make an image using ddrescue Now two weeks later I have an image of the luks partition in a .img file.

# ls -lh
total 14T
-rw-r--r-- 1 root root 14T Oct 25 01:57 candybox.img
-rw-r--r-- 1 root root 271 Oct 20 14:32 candybox.logfile

After numerous attempts using everything I could find online I could not coerce e2fsck to do anything on the image, so I used mkfs.ext4 -L candy candybox.img -m 0 -S and I was able to mount the dirty filesystem readonly without the journal and recover 960G of data. It gave all kinds of errors of various directories not existing and so forth but I was able to get some stuff. Which gave me some hope!

I then ran e2fsck again and it had to recreate the root inode and gave a massive list of correcting group counts, I accepted the root inode creation and said no to everything else, leaving a completely empty filesystem. Re-ran again and said yes to all questions with the same result but now a "clean" but empty filesystem.

extundelete gives me 0 recoverable inodes found.

And now I'm stuck again, I can't come up with any other methods other than dropping to something like photorec which will give me an absolute mess with how large the filesystem was.

I'm willing to re-copy the image from the original array and start over, if I can get any suggestions or ideas on a way to get more of my files back.

I wish I could give more detailed logs of the commands that have run, but the output is long scrolled passed except for what gets logged to syslog and my memory is not as detailed due to the timeframe this has occurred over.

Any help is greatly appreciated!

Update Oct 27

I've fully recopied the image to start testing on again, and here is the output so far. The copy process:

[root@gamma rescue]# nbd-client 172.16.10.204 2000 /dev/nbd0
Negotiation: ..size = 14307292MB
bs=1024, sz=15002283540480 bytes
[root@gamma rescue]# cryptsetup luksOpen /dev/nbd0 candybox
Enter passphrase for /dev/nbd0: 
[root@gamma mnt]# pvcreate /dev/md5
  Physical volume "/dev/md5" successfully created
[root@gamma mnt]# pvscan
  PV /dev/md5                      lvm2 [18.19 TiB]
  Total: 1 [18.19 TiB] / in use: 0 [0   ] / in no VG: 1 [18.19 TiB]
[root@gamma mnt]# vgcreate vg-rescue /dev/md5
  Volume group "vg-rescue" successfully created
[root@gamma mnt]# lvcreate --size 15T --name lv-rescue vg-rescue
  Logical volume "lv-rescue" created
[root@gamma mnt]# mkfs.xfs /dev/vg-rescue/lv-rescue 
log stripe unit (524288 bytes) is too large (maximum is 256KiB)
log stripe unit adjusted to 32KiB
meta-data=/dev/vg-rescue/lv-rescue isize=256    agcount=33, agsize=125828992 blks
         =                       sectsz=512   attr=2
data     =                       bsize=4096   blocks=4026531840, imaxpct=5
         =                       sunit=128    swidth=640 blks
naming   =version 2              bsize=4096   ascii-ci=0
log      =internal log           bsize=4096   blocks=521728, version=2
         =                       sectsz=512   sunit=8 blks, lazy-count=1
realtime =none                   extsz=4096   blocks=0, rtextents=0
[root@gamma mnt]# mount /dev/vg-rescue/lv-rescue rescue/
[root@gamma rescue]# ddrescue /dev/mapper/candybox candybox.img candybox.ddlog


Press Ctrl-C to interrupt
Initial status (read from logfile)
rescued:         0 B,  errsize:       0 B,  errors:       0
Current status
rescued:    13194 GB,  errsize:   1807 GB,  current rate:        0 B/s
   ipos:    13194 GB,   errors:       1,    average rate:   73528 kB/s
   opos:    13194 GB,     time from last successful read:      44 s
^Clitting failed blocks... 
Interrupted by user
## Network hung, had to try again here
[regan@gamma ~]$ sudo nbd-client -d /dev/nbd0
Disconnecting: que, disconnect, Error: Ioctl failed: Invalid argument

Exiting.
[regan@gamma ~]$ sudo nbd-client 172.16.10.204 2000 /dev/nbd0
Negotiation: ..size = 14307292MB
bs=1024, sz=15002283540480 bytes

[root@gamma rescue]# ddrescue -r 2 /dev/mapper/candybox candybox.img candybox.ddlog


Press Ctrl-C to interrupt
Initial status (read from logfile)
rescued:    15002 GB,  errsize:   7426 kB,  errors:      60
Current status
rescued:    15002 GB,  errsize:       0 B,  current rate:    77529 kB/s
   ipos:    15002 GB,   errors:       0,    average rate:    69297 kB/s
   opos:    15002 GB,     time from last successful read:       0 s
Finished                       

[root@gamma rescue]# lvcreate -l 100%FREE -s -n rescue_snap /dev/vg-rescue/lv-rescue 
  Logical volume "rescue_snap" created
[root@gamma rescue]# cd ..
[root@gamma mnt]# mount -o remount,ro rescue/
[root@gamma mnt]# mkdir rescue_snap
[root@gamma mnt]# mount -o nouuid /dev/vg-rescue/rescue_snap rescue_snap
[root@gamma mnt]# cd rescue_snap/
[root@gamma rescue_snap]# ls
candybox.ddlog  candybox.img

The messy:

[root@gamma rescue_snap]# mkfs.ext4 -L candy candybox.img -m 0 -S
mke2fs 1.41.10 (10-Feb-2009)
candybox.img is not a block special device.
Proceed anyway? (y,n) y
Filesystem label=candy
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
Stride=0 blocks, Stripe width=0 blocks
915668992 inodes, 3662666368 blocks
0 blocks (0.00%) reserved for the super user
First data block=0
Maximum filesystem blocks=4294967296
111776 block groups
32768 blocks per group, 32768 fragments per group
8192 inodes per group
Superblock backups stored on blocks: 
    32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 
    4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968, 
    102400000, 214990848, 512000000, 550731776, 644972544, 1934917632, 
    2560000000

Skipping journal creation in super-only mode
Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 26 mounts or
180 days, whichever comes first.  Use tune2fs -c or -i to override.

[root@gamma rescue_snap]# mount -o loop candybox.img /mnt2
[root@gamma rescue_snap]# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/md2              147G  138G  3.1G  98% /
tmpfs                  16G  5.7M   16G   1% /dev/shm
/dev/md0              494M  199M  276M  42% /boot
/dev/sdc1             1.8T  979G  763G  57% /mnt/macmirror
/dev/sdj1             1.8T  970G  771G  56% /mnt/usbrescue
/dev/mapper/vg--rescue-lv--rescue
                       15T   14T  1.4T  91% /mnt/rescue
/dev/mapper/vg--rescue-rescue_snap
                       15T   14T  1.4T  91% /mnt/rescue_snap
/mnt/rescue_snap/candybox.img
                       14T   15M   14T   1% /mnt2

## Even though it says only 15M is used, I was able to rsync 960G to /mnt/usbrescue

[root@gamma rescue_snap]# cd /mnt2/
[root@gamma mnt2]# ls -l
ls: cannot access Fedora-19-x86_64-DVD: Input/output error
ls: cannot access rsync_batch: Input/output error
ls: cannot access shell1: Input/output error
ls: cannot access New Folder (2): Input/output error
ls: cannot access shell2: Input/output error
ls: cannot access revolution: Input/output error
ls: cannot access mail: Input/output error
ls: cannot access testing: Input/output error
ls: cannot access export: Input/output error
ls: cannot access ben_backup_20130903: Input/output error
total 160488672
drwxr-xr-x     2 regan regan        4096 Sep  3 20:16 100MEDIA
drwxr-xr-x    19 regan regan        4096 Sep 26 05:18 android
d??????????    ? ?     ?               ?            ? ben_backup_20130903
-rw-rw-r--     1 regan regan       12126 Jan  4  2013 durations.txt
d??????????    ? ?     ?               ?            ? export
drwxrwxr-x    10 regan regan        4096 Dec 29  2012 family-pc_20121229
d??????????    ? ?     ?               ?            ? Fedora-19-x86_64-DVD
-rw-r--r--     1 regan regan 72116729363 Sep 30 04:39 gamma_backup_20130928.tgz
-rw-rw-r--     1 regan regan 55606528323 Jul 27  2011 gamma_tar_20110727.tbz2
-rw-rw-r--     1 regan regan        3839 Sep 27  2012 Good Quality2.plist
-rw-rw-r--     1 regan regan        4663 Oct  7  2012 Good Quality3.plist
-rw-rw-r--     1 regan regan        3852 Sep 26  2012 Good Quality.plist
drwxr-xr-x     7 regan regan        4096 Nov 13  2012 grok
d??????????    ? ?     ?               ?            ? HardDisks
-rwxr--r--     1 regan regan       54248 Mar 16  2013 IMAG0868.jpg
-rwxr--r--     1 regan regan       51156 Mar 16  2013 IMAG0869.jpg
-rwxr--r--     1 regan regan       85912 Mar 16  2013 IMAG0870.jpg
-rwxr--r--     1 regan regan       76875 Mar 16  2013 IMAG0872.jpg
-rwxr--r--     1 regan regan       68451 Mar 16  2013 IMAG0873.jpg
-rwxr--r--     1 regan regan       59587 Mar 16  2013 IMAG0874.jpg
-rwxr--r--     1 regan regan       81232 Mar 16  2013 IMAG0875.jpg
-rwxr--r--     1 regan regan       44211 Mar 16  2013 IMAG0876.jpg
-rwxr--r--     1 regan regan       41660 Mar 16  2013 IMAG0877.jpg
-rwxr--r--     1 regan regan       36778 Mar 16  2013 IMAG0878.jpg
-rwxr--r--     1 regan regan       76964 Mar 16  2013 IMAG0879.jpg
-rwxr--r--     1 regan regan       81876 Mar 16  2013 IMAG0880.jpg
-rwxr--r--     1 regan regan     1568002 Mar 16  2013 IMAG0953.jpg
-rwxr--r--     1 regan regan     1548566 Mar 16  2013 IMAG0954.jpg
-rwxr--r--     1 regan regan     1351743 Mar 16  2013 IMAG0955.jpg
-rwxr--r--     1 regan regan     1750128 Mar 16  2013 IMAG0956.jpg
-rwxr--r--     1 regan regan     1694378 Mar 16  2013 IMAG0957.jpg
-rwxr--r--     1 regan regan     1277128 Mar 16  2013 IMAG0958.jpg
-rwxr--r--     1 regan regan     1467452 Mar 16  2013 IMAG0965.jpg
-rwxr--r--     1 regan regan     1595903 Mar 16  2013 IMAG0966.jpg
-rwxr--r--     1 regan regan     1372444 Mar 16  2013 IMAG0967.jpg
-rwxr--r--     1 regan regan     1698010 Mar 16  2013 IMAG0968.jpg
-rwxr--r--     1 regan regan     1550641 Mar 16  2013 IMAG0969.jpg
-rwxr--r--     1 regan regan     1333768 Mar 16  2013 IMAG0970.jpg
-rwxr--r--     1 regan regan     1432347 Mar 16  2013 IMAG1010.jpg
-rwxr--r--     1 regan regan     1668159 Mar 16  2013 IMAG1013.jpg
-rwxr--r--     1 regan regan     1657058 Mar 16  2013 IMAG1014.jpg
-rwxr--r--     1 regan regan     1496547 Mar 16  2013 IMAG1016.jpg
-rwxr--r--     1 regan regan     1609156 Mar 16  2013 IMAG1017.jpg
-rwxr--r--     1 regan regan     1604832 Mar 16  2013 IMAG1019.jpg
-rwxr--r--     1 regan regan     2048916 Mar 16  2013 IMAG1073.jpg
-rwxr--r--     1 regan regan     2006024 Mar 16  2013 IMAG1074.jpg
-rwxr--r--     1 regan regan     1926686 Mar 16  2013 IMAG1075.jpg
-rw-r--r--     1 regan regan     1583090 Jul 14 21:15 IMAG1565.jpg
-rw-r--r--     1 regan regan     1435031 Sep 22 05:19 IMAG1762.jpg
-rw-r--r--     1 regan regan     1531602 Sep 22 05:19 IMAG1763.jpg
-rw-r--r--     1 regan regan     1450926 Sep 22 05:19 IMAG1764.jpg
-rw-r--r--     1 regan regan     1336103 Sep 23 21:31 IMAG1765.jpg
-rw-r--r--     1 regan regan     1235885 Sep 23 21:32 IMAG1766.jpg
-rw-r--r--     1 regan regan     1224376 Sep 23 21:32 IMAG1767.jpg
-rw-r--r--     1 regan regan     1235229 Sep 23 21:32 IMAG1768.jpg
drwxrwxr-x     2 regan regan        4096 Mar  9  2013 jakeanmal
-rw-rw-rw-     1 regan regan      115228 Oct 29  2009 jj_nas
drwx------.    2 root  root        16384 Nov  8  2012 lost+found
-rw-r--r--     1 regan regan  3123877728 Nov  6  2010 luridmirror_20090806.tar.xz
-rw-r--r--     1 regan regan  2877033943 Mar  1  2013 macabre_20130301.tgz
d??????????    ? ?     ?               ?            ? mail
-rw-r--r--     1 root  root         6771 Aug 10  2009 mail_mirror
-rw-------     1 regan regan 21913047552 Apr  4  2013 mallorys_hdd.vbox.img
d??????????    ? ?     ?               ?            ? MSDN
-rw-r--r--     1 regan regan        8572 May 10  2010 Music
d??????????    ? ?     ?               ?            ? New Folder (2)
drwxrwxrwx    24 regan regan        4096 Mar 22  2013 onyx
drwxr-xr-x   231 regan regan       24576 Sep 30 09:29 ptp
-rwxr--r--     1 regan regan      483328 Jan 26  2013 putty.exe
d??????????    ? ?     ?               ?            ? revolution
-rw-r--r--     1 root  root   6272757760 Oct 16  2012 root.tar
d??????????    ? ?     ?               ?            ? rsync_batch
drwxrwxr-x     2 regan regan       12288 Oct  6 04:09 saber
drwxrwxr-x     2 regan regan      188416 Sep 25 04:20 session_tmp
d??????????    ? ?     ?               ?            ? shell1
d??????????    ? ?     ?               ?            ? shell2
d??????????    ? ?     ?               ?            ? testing
drwxrwxr-x     3 regan regan        4096 Oct  7  2012 tofix
-rwxr--r--     1 regan regan    64991966 Jan  2  2013 VIDEO0041.mp4

[root@gamma mnt2]# cd ..
[root@gamma /]# umount /mnt2
[root@gamma /]# cd /mnt/rescue_snap/
[root@gamma rescue_snap]# e2fsck candybox.img 
e2fsck 1.41.10 (10-Feb-2009)
Backing up journal inode block information.

candy contains a file system with errors, check forced.
Resize inode not valid.  Recreate<y>? yes

Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Root inode not allocated.  Allocate<y>? yes

/lost+found not found.  Create<y>? yes

Pass 4: Checking reference counts
Pass 5: Checking group summary information
Block bitmap differences:  -(9252--9255) <Snip a few dozen MB of text> -(3662666237--3662666238) -3662666240 -(3662666242--3662666244) -(3662666247--3662666249) -3662666253 -(3662666255--3662666256) -3662666259 -3662666262 -3662666264 -(3662666268--3662666271) -3662666276 -3662666281 -3662666285 -3662666294 -(3662666296--3662666297) -3662666301 -3662666307 -3662666309 -3662666311 -3662666313 -3662666316 -(3662666318--3662666319) -3662666324 -(3662666326--3662666328) -(3662666331--3662666332) -3662666334 -(3662666341--3662666342) -3662666344 -(3662666346--3662666347) -3662666349 -(3662666351--3662666352) -3662666354 -3662666357 -3662666362 -(3662666366--3662666367)
Fix<y>? yes

Free blocks count wrong for group #0 (23517, counted=23516).
Fix<y>? 

Free blocks count wrong (3605188902, counted=3605188901).
Fix<y>? yes

Free inodes count wrong for group #0 (8191, counted=8190).
Fix<y>? yes

Directories count wrong for group #0 (1, counted=2).
Fix<y>? yes

Free inodes count wrong (915668991, counted=915668990).
Fix<y>? yes


candy: ***** FILE SYSTEM WAS MODIFIED *****
candy: 2/915668992 files (0.0% non-contiguous), 57477467/3662666368 blocks
[root@gamma rescue_snap]# mount -o loop candybox.img /mnt2
[root@gamma rescue_snap]# ls -l /mnt2
total 4
drwx------ 2 root root 4096 Oct 27 19:33 lost+found
[root@gamma rescue_snap]# 

Note that I now have my backup image in a snapshot, so I can try theories over and over if anyone has some ideas...

Solution

My preferred tool for filesystem data recovery is UFS Explorer.

I do not know how the LUKS encryption will impact this, but download an evaluation of UFS Explorer and try the normal best-practices. Run it against your image and see if things appear to be more recoverable. The tool will provide you with a good view of the directory structure and give some ability to search for data.

Also see: How to recover XFS file system with "superblock read failed"

Comments

[MadHatter]

Regan, welcome to SF! May I congratulate you on a well-written, pertinent, well-researched question? Sadly, I'm not enough of an ext guru to give you any more advice - you've already done everything I would have tried. I very much hope that someone with more clue than me can give you some hints about where to go from here.

[charlesbridge]

What happened to the original 2T drives that you pulled from the array? Did you destroy them already?

[Regan]

I've used them for other things already, as I removed them from the array, once it successfully rebuilt I then used the 2T drive in something else. Currently they are being used as part of my /mnt/rescue raid0. I wish they were untouched, as I could have used them to recover from.

[Michael Hampton]

Where are your backups?

[ewwhite]

You mentioned LUKS. Is/was this encrypted on-disk as well?

[poige]

"Mr. Sigmund Freud tried everything and what he didn't liked he called a perversion…" ©

[Regan]

MichaelHampton: I've never considered backing up this volume before as I originally used it to store backups, but that changed overtime and now I have some new planning to do. @ewwhite: Yes, it is encrypted. It's ext4 over luks over md raid. So I mdadm -assemble, luksopen /dev/md0, then I mount /dev/mapper/candybox. The raid array is fully intact, the luks volume opens flawlessly, but e2fsck -y wreaked havoc when I ran it. All data was accessible before umount; e2fsck.

[Grant]

No idea about fixing it but when you do I would advise not using raid5 on such large volumes. It takes a long time to recover and there is a large chance of another disk dying.

[Regan]

@Grant thanks, I've discovered that myself already. Once I'm in "success" or "give up" stage, I plan to rebuild my volume from the ground up with things I've learned the hard way. Including automated backup for sure... What advice do you have for such a large volume, assuming you have experience dealing with one?

[Grant]

@Regan I have used RAID6 for several 10+TB arrays. Since it can withstand any 2 drives failing chances to rebuild are much better. For one system instead of raid I have used greyhole which simply keeps multiple copies of each file across multiple drives. Set to two copies it means losing one drive will lose no data. Losing two drives will lose some - but not all - data. And always have backups of anything you cant easily replace.

[Regan]

Keep in mind, the volume I'm trying to recover is ext4. I used XFS on my rescue volume because my volume was originally 18T and 32bit ext4 didn't want to run (limit of 16T with standard blocksize). I will give UFS explorer a try though and see what happens

[ewwhite]

UFS Explorer works fine for ext2, ext3 and ext4 systems.

[Regan]

UFS Explorer wants an eta of 41 days to scan the volume... I'm going to attempt the debian version tomorrow, maybe it'll result faster than winblows.

[ewwhite]

@Regan Did it work for you?

[Regan]

It did not, it wasn't able to see the ext4 filesystem at all after I installed a linux version and let it deep scan the entire volume. I've emailed the ext4 dev mailing list and they've been sending me ideas trying to narrow down the cause. If I get an answer that works from them I'll post it up. -- I think UFS may work to recover a different issue a friend of mine has, glad you pointed me towards it, it's a pretty sweet tool!