Well my old hard drive seems to have partially failed right out of the blue.

doing a normal mount complains about superblocks.

mount /dev/sdb1 /mnt/sdb1

I try to backup the device fist using "dd" but that starts to copy then quickly fails, only copying about 1.2 megs.

I then try using "fsck" and "e2fsck", I get the error message:

Attempt to read block from filesystem resulted in short read while trying to open /dev/sdb1 Could this be a zero-length partition?

Using "mke2fs -n /dev/sdb1" I do get some information out.


If I use "cfdisk" or "fdisk" it complains about not being able to read the device.


so I rebooted my machine and yes it suddenly takes a bit longer for the old hard drive to be read, so indicating some hardware issue, but it does eventually read it.

As it had an old installation of Linux on there, I tried booting into it, it did start to boot up but then failed to mount it self.


I am guessing the circuit board on the hard disk is faulty and probably if I can source an identical copy I could replace it. The disk wasn't making any noises so I don't think it's an internal issue, currently I have it unplugged just to preserve it.


The data on there isn't that important but some of it is a bit of a shame, stuff that I hadn't yet backed up.


So are there any other tricks?

I doubt it but would an external sata to usb connector work?

or do you think the circuit board would need replacing?

If your hard disk failed just remove it.To remove /dev/sdb, we will mark /dev/sdb1 and /dev/sdb2 as failed and remove them from their respective RAID arrays.

I am not using a RAID

Everything you say sounds like hardware failure. If you want to try to recover the data you might try Testdisk. Good Luck

I know the hardware issue might be a number of things, but as the hard drive wasn't making any nasty noises, it sounded as if it was still spinning correctly and the fact it still attempted to boot the device.

Is it quite likely that just getting a matching donor Hard drive and swapping the circuit board a likely solution? or could there be a more serious internal hardware issue?


However I can guess if I do replace the circuit board I still have to do some commands to get the data off due to the drive not closing correctly (although the drive was set up as read only when it failed).

Again I would like to know if it is a good chance that its just a faulty circuit board before I go on a wild goose chase and try to fix it.

Probably not, but...

If no live-cd sees the files and "fdisk" does not see the partitions, but the HDD spins, it is hard to know what is wrong.

See if the HDD maker has a HDD utility program you can download, burn and run. There is HUTIL, SeaTools, etc. It might give you more information about the fault.

Do you have SMART activated in your BIOS and on the HDD? You could try that, again to gain more information.

Have you tried putting the HDD in another system to be sure the problem is the HDD?

A fault in the HDD as a device is a good guess, but if the problem were the circuit board, would you not expect that the HDD would give you no response at all, that is the condition would be one of off/on, no-spin/spin?

From my own experience, when the circuit board on a HD dies, the drive won't even spin-up. If the drive is spinning up, then it sounds like the disk is going bad. You can backup whatever you can get off of it with the following command:

The noerror will tell dd not to quit if it hits a read error (so it'll get past the initial 1.2 mb), and sync will fill in the bad areas with null characters.

You should get whatever data you can off of it, then stop using it. Usually, when a drive starts dying, bad sectors start multiplying very rapidly.

thanks for that "conv=noerror,sync" bit, I am currently ""dding" it using that (this will take a while, curse large hard drives). Well once it has finish, I'll try stuff I have read elsewhere. See if I can recover some of the rubbish I have

Not a problem. I had to go through that whole mess once to recover the actual audio recordings off an old voicemail system that had the hard drive start to die. After you get the image, you can use testdisk to try and recover the partition data.

well I'll see how it goes, I would be amazed if I can't recover anything.

I think I will be amazed. I haven't finished dding but its just complaining of input/output error all the time and viewing the iso with less just seems filled of @ symbols and can't see anything else at all.


I think I may have lost this battle.

Don't feel too discouraged, most binary characters will just show up as @ in less. If you want to try and speed it up, set a blocksize with bs=1024 or bs=2048. The default is 512 bytes. The pro of having a higher block size is it will run faster. The con is when it hits an error, it skips over the entire block.

It may also be that the beginning of your drive is in really bad shape, but it may be better off later on. Either way, having a hard drive crash is never fun.

sadly didn't seem to work, just filled the entire iso with null characters, running "ddrescue" just failed to read anything.

It appears that whatever the fault is, its just not actually reading anything at all, as in closing your eyes and looking at a book (there is stuff there, but just not seeing it).

personally never seen a hdd fail like this before, even in a head crash you can get data off via dd.

oh well only can think of 1 folder that I forgot to backup. So its not a great loss.