Introduction
This guide is for ext4 , though I’ll explain how other filesystems can be cured along the way. The easiest way to carry all this out, seeing as your computer probably won’t boot at this stage, is to download and burn a copy of Parted Magic. Boot from that, and you’ll get access to a number of useful tools.
The procedure
First, figure out what partition we’re dealing with.
sudofdisk-l
The above will list all the partitions on all the drives in your computer. To recover a lost partition, you’re going to need Testdisk. Testdisk is included in Parted Magic, and there’s a great guide on their site. For this though, we just need the partition number, such as /dev/sda3 or /dev/hdb1.
Now, make sure your superblock is the problem, by starting a filesystem check, replacing xxx with your partition name. Here, you can change ext4 to ext3, or ext2 to suit the filesystem.
sudofsck.ext4 -v/dev/xxx
If your superblock is corrupt, the output will look like this
fsck /dev/sda5
fsck 1.41.4 (27-Jan-2009)
e2fsck 1.41.4 (27-Jan-2009)
fsck.ext4: Group descriptors look bad... trying backup blocks...
fsck.ext4: Bad magic number in super-block while trying to open /dev/sda5
The superblock could not be read or does not describe a correct ext4
filesystem. If the device is valid and it really contains an 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>
Now lets find where your superblock backups are kept.
sudomke2fs -n /dev/xxx
Down at the bottom of this output, should be a list of the backups:
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208
You’re almost there. Finally, restore the superblock from the backup, again replacing the x’s with your partition name, and block_number with the first backup superblock.
sudoe2fsck -b block_number /dev/xxx
Now reboot, and your superblock should be fixed. If it’s not, repeat the steps, but restore a different backup superblock 
Checking your file system as a precaution
If you want to check your Pi file system because you feel it might be having a problem, try the following:
First of all, shut down your Pi using the following which will make the Pi check it’s own file system on reboot:
sudo shutdown -F -r now
Once the Pi comes up again, check the results of the scan by looking at the following file:
/var/log/fsck/
[…] recantha Updated 6th March 2013 with a way to check your file system as a precaution (see the bottom of this […]
All I get is:
e2fsck: Invalid argument while trying to open /dev/sdx1
At what point do you get the error?
After I try to restore the superblock with sudo e2fsck -b block_number my sd card.
Try a different superblock.
I tried them all.
I don’t know what to suggest, then – it’s always worked for me and others 🙁
Same, I tried all the superblock backups. I get the same error. “Superblock could not be read or does not describe a valid ext2/ext3/ext4 filesystem”
In that case I would say your SD card is beyond repair. 🙁
Hi Rob, I had the exactly same problem. I managed to fix it by inserting the sd card in windows and deleted the cmdline.txt with corrupted content at the boot partition. Then Ive powered up the pi with this card again and the OS took care of the damaged filesystem by itself.
[…] 首先你应该做的是拷贝一份你的SD卡镜像，把你的SD卡插入Linux系统的电脑，然后使用dd命令来拷贝你的镜像文件，然后你可以在不进一步破坏数据的情况下修复这个镜像文件。你应该做的第一件事儿就是使用备份文件来替换你的SD卡超级块，具体步骤可参考这里。如果还是不起作用，一切都失败了，可以使用自定义的恢复配置文件来救回你的文件，如何操作可以参考这里。 […]
Use Win Minitool Power Data Recovery i better.
I followed your instructions to repairing a broken CD superblock on RASPi, it does not appear to be taking the command. Below is my output.
$ sudo fsck.ext4 -v /dev/sdb1
e2fsck 1.42.9 (4-Feb-2014)
ext2fs_open2: Bad magic number in super-block
fsck.ext4: Superblock invalid, trying backup blocks…
fsck.ext4: Bad magic number in super-block while trying to open /dev/sdb1
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
or
e2fsck -b 32768
$ sudo mke2fs -n /dev/sdb1
mke2fs 1.42.9 (4-Feb-2014)
Filesystem label=
OS type: Linux
Block size=1024 (log=0)
Fragment size=1024 (log=0)
Stride=0 blocks, Stripe width=0 blocks
14336 inodes, 57344 blocks
2867 blocks (5.00%) reserved for the super user
First data block=1
Maximum filesystem blocks=58720256
7 block groups
8192 blocks per group, 8192 fragments per group
2048 inodes per group
Superblock backups stored on blocks:
8193, 24577, 40961
$ sudo e2fsck -b 8193 /dev/sdb1
e2fsck 1.42.9 (4-Feb-2014)
e2fsck: Invalid argument while trying to open /dev/sdb1
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
or
e2fsck -b 32768
I did the same for all three backups of the supperblock with the same results.
Interestly when I launch Gparted and look at the SD card is it showing for the boot partition is a Fat16. is there away to change or correct it without loosing my data.
I was using my raspi as database server mainly and I did not have a backup before the corruption occurred.
Please advise.
Thanks in advanced.
theAntiguan
I think probably the best thing to do is to post your question to the Raspberry Pi Forum – it sounds like you’re having a very specific problem and I’m afraid the level of my knowledge might set you on the wrong path.
I had exactly this problem on several new µSD.
In fact the problem was related to a very beautiful Samsung adapter !
When I came back to my usual adapter, everything was correct.
Best regards
mount the boot partition in another machine, delete cmdline.txt then power up pi again
I have the same problem.
“The superblock could not be read or does not describe a correct ext2 filesystem. If the device is valid and it really contains an ext2 filesystem, then the superblock is corrupt, and you might trey running e2fsck with an alternate superblock: e2fsck -b 8193 ”
I use all the blocks but then again it shows the same problem.
Do you have any idea how can i solve this?
Thank you!!!
Elisavet
No, sorry, I don’t – it normally works for me. Try posting to the Raspberry Pi Forums – there’s more people to possibly help you on there 🙂
Hi Michael, just wanted to say thanks for the excellent post here. I had issues when power was lost to a Pi acting as a wallboard monitor for one of our company’s websites. The Pi refused to boot and all the code behind the monitoring was (foolishly on my side…) on the Pi only and not backed up anywhere.
Thank you very much! Seems like this solution worked for me.
Now I’m looking for the best solution to prevent this to happen again.
Maybe a job in the boot process with the command:
sudo fsck.ext4 -v /dev/xxx
Regards.
I was able to recover several days of work by repairing mi pi’s SD card following your instructions.
Thank you!
You’re very welcome indeed 🙂
Thank you very much for your fantastic howto!! You saved me a lot of work. Thanks again, have a great day (-:
You’re welcome. Glad it helped 🙂
Thank you for your help. My question is, do you have any idea what can be the origin of this problem? First I thought in power failures, but I could not find any crash with last command.
I’ve generally seen it happen when you pull the power out of the Pi without shutting it down properly.
Yes, that was my first idea, but I could not detect none. I use “last” command, which prints “crash” whenever a power failure happens (as far as I understood, searching for how could I detect power failures from the logs), but none occurred. Do you know other way to detect if a power failure actually happened?
As far as I know, you don’t get any logging when a power failure happens because the computer is ‘off’ before it can log anything. I don’t know of any other method… except the SD card normally corrupts (even if it’s in a small way).
If you need to track power outages, the only thing I can suggest is a cron job that logs something every minute with a timestamp and then you can check for gaps.
This last suggestion about doing Shutdown with -r -F didn’t work. On reboot there was no directory fsck.
To be honest, this post is getting a bit old, so it might need updating!