fsck raid10 ext4 partition currupted, restore data

Question

there is a volume in raid10 /dev/md3 it has a GPT partition /dev/md3p1 in ext4 with a size of 16TB.

I accidentally ran

fsck -y /dev/md3

causing the filesystem in /dev/md3p1 to get corrupted.

fsck -b superblock_number /dev/md3p1 with any superblock from mkfs -n /dev/md3p1 returns an error.

testdisk in /dev/md3 sees ext4 partition, but in /dev/md3p1 sees nothing.

How to recover data or partition?

sudo parted -l /dev/md3

Model: ATA WDC WD8003FFBX-6 (scsi)
Disk /dev/sda: 8002G
Sector size (logical/physical): 512B/4096B
Partition Table: gpt
Disk Flags: 

Number  Start   End     Size    File system  Name  Flags
 1      1049kB  538MB   537MB   fat32              boot, esp
 2      538MB   1612MB  1074MB
 3      1612MB  55.3GB  53.7GB
 4      55.3GB  66.0GB  10.7GB
 5      66.0GB  8002GB  7936GB


Model: ATA WDC WD8003FFBX-6 (scsi)
Disk /dev/sdb: 8002GB
Sector size (logical/physical): 512B/4096B
Partition Table: gpt
Disk Flags: 

Number  Start   End     Size    File system  Name  Flags
 1      1049kB  538MB   537MB   fat32              boot, esp
 2      538MB   1612MB  1074MB
 3      1612MB  55.3GB  53.7GB
 4      55.3GB  66.0GB  10.7GB
 5      66.0GB  8002GB  7936GB


Model: ATA WDC WD8003FFBX-6 (scsi)
Disk /dev/sdc: 8002GB
Sector size (logical/physical): 512B/4096B
Partition Table: gpt
Disk Flags: 

Number  Start   End     Size    File system  Name  Flags
 1      1049kB  538MB   537MB   fat32              boot, esp
 2      538MB   1612MB  1074MB
 3      1612MB  55.3GB  53.7GB
 4      55.3GB  66.0GB  10.7GB
 5      66.0GB  8002GB  7936GB


Model: ATA WDC WD8003FFBX-6 (scsi)
Disk /dev/sdd: 8002GB
Sector size (logical/physical): 512B/4096B
Partition Table: gpt
Disk Flags: 

Number  Start   End     Size    File system  Name  Flags
 1      1049kB  538MB   537MB   fat32              boot, esp
 2      538MB   1612MB  1074MB
 3      1612MB  55.3GB  53.7GB
 4      55.3GB  66.0GB  10.7GB
 5      66.0GB  8002GB  7936GB

sudo mount /dev/md3p1 /1

mount: /1: wrong fs type, bad option, bad superblock on /dev/md3p1, missing codepage or helper program, or other error.

sudo mkfs -n /dev/md3p1



mke2fs 1.44.1 (24-Mar-2018)
Creating filesystem with 3874701824 4k blocks and 484339712 inodes
Filesystem UUID: dc34be44-6f5a-47ac-a841-cbb75625b9b0
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, 3855122432

sudo fsck -b 32768 /dev/md3p1

fsck from util-linux 2.31.1
e2fsck 1.44.1 (24-Mar-2018)
fsck.ext2: Bad magic number in super-block while trying to open /dev/md3p1

The superblock could not be read or does not describe a valid ext2/ext3/ext4
filesystem.  If the device is valid and it really contains an ext2/ext3/ext4
filesystem (and not swap or ufs or something else), then the superblock
is corrupt, and you might try running e2fsck with an alternate superblock:

    e2fsck -b 8193 <device>

 or

    e2fsck -b 32768 <device>

cat /proc/mdstat


Personalities : [raid1] [raid10] [linear] [multipath] [raid0] [raid6] [raid5] [raid4] 
md0 : active raid1 sdc2[6] sdb2[5] sda2[4] sdd2[0]
      1046528 blocks super 1.2 [4/4] [UUUU]
      
md1 : active raid1 sdb3[6] sdc3[5] sdd3[0] sda3[4]
      52395008 blocks super 1.2 [4/4] [UUUU]
      
md2 : active raid10 sdb4[5] sda4[4] sdc4[3] sdd4[0]
      20953088 blocks super 1.2 512K chunks 2 near-copies [4/4] [UUUU]
      
md3 : active raid10 sda5[1] sdd5[0] sdc5[3] sdb5[2]
      15498811392 blocks super 1.2 512K chunks 2 near-copies [4/4] [UUUU]
      bitmap: 0/116 pages [0KB], 65536KB chunk

Please indicates exactly what the error messages are. Also add the output from `parted -l /dev/md3`. — wazoox, Dec 23 '20 at 15:28
yes, hot shutdown, after accidentally ran fsck -y /dev/md3 instead of fsck -y /dev/md3p1 — user607743, Dec 27 '20 at 09:40

score 1 · Accepted Answer · answered Dec 28 '20 at 20:54

1

if you have tried with all superblocks, you may want to check out mke2fs' man mage, and maybe try it with the -S to re-init superblock and group descriptors (first, see the man page for warnings).

... you need to run fsck immediately afterwards, and have the right block size in hand.

answered Dec 28 '20 at 20:54

JRM.

26
1

mkfs -S /dev/md3p1 ? and fsck -b block_number /dev/md3p1 – user607743 Dec 29 '20 at 10:32
correct, since mkfs calls mke2fs. Please check the man page before you attempt it: https://linux.die.net/man/8/mke2fs ...and again, I would try this only if you have already tried using all the superblock backups. – JRM. Dec 29 '20 at 17:01
...also, know that after running fsck, there might be some stuff dumped into lost+found. ...so again, I would try this only if you have already tried using all the superblock backups. – JRM. Dec 29 '20 at 17:15
mkfs -S /dev/md3p1 ...or if you happen to know the block size of the partition: mkfs -S -b /dev/md3p1 ...maybe this can help getting the block size: dumpe2fs /dev/md3p1 | grep -i block | grep -i size or stat -f /dev/md3p1 dumpe2fs will also get you all the superblocks backups listed,.. then you can make sure mkfs -n list matches dumpe2fs. – JRM. Dec 29 '20 at 18:10

fsck raid10 ext4 partition currupted, restore data

1 Answers1