First, I used the exact commands in the related Ubuntu Wiki:
https://help.ubuntu.com/community/DataRecovery . I installed gddrescue and followed the instructions.
After that, I have used the commands in a post of Anandtech forums:
http://forums.anandtech.com/showpost...02&postcount=7
However, mdadm command didn't work in my case. I got the "superblock problem". Then, I used
Code:
mdadm --build --level=0 --raid-devices=2 --chunk=64 /dev/md0 /dev/loop0 /dev/loop1
I have RAID0 (level=0) and 2 hard drives -therefore 2 hdd images- (raid-devices=2). You need to specify the chunk size (chunk=64), or you cannot mount the partitions under RAID0. I have used the RAID BIOS to determine chunk size (it was written on the screen when I entered the RAID utility, using CTRL+I).
/dev/md0 is the block device defines the RAID array. You must use an unused md device to build the RAID array.
/dev/loop0 is the image of the first hdd
/dev/loop1 is the image of the second hdd
In your case, you may destroy your image or your RAID array completely but I was desperate and I needed to do something to rescue only 1 file, then acted bravely. I also took backups of the image files to some other storage media, because taking the images from the broken disks lasted 5 hours.
Also, you can use dmraid but I couldn't manage to define the loopback devices as an input to dmraid. Automatic find didn't work in my case.
It is a bit complex and as I said you may destroy your data. Please do not take this explanation as the ultimate solution, try the solutions on Ubuntu Wiki and Anandtech forums at first, then look at the man pages of mdadm and try the above solution as your final move.
The info about the mdadm package I used:
Code:
# aptitude show mdadm
Package: mdadm
State: installed
Automatically installed: no
Version: 3.2.5-1ubuntu3
Note: I have Googled a lot, tried a lot and found that the info described in the links are much more better than the others. I didn't want to credit everything to me and I gave the links. If this is against the rules, I am sorry.