When a drive is failing, the safe order is: image it first, then work only on the image.
possibly outdated More than two years have passed since publication. The content may no longer hold.
1. Image the drive
dd_rescue /dev/damaged_hdd /mountpoint/file
2. Find the partitions
Attach the image to a loop device and print its partition table:
losetup /dev/loop0 /mountpoint/file fdisk -l -u /dev/loop0
Note the start and end sectors of the partition you need (for example a Linux raid autodetect partition).
Then map just that partition to its own loop device:
losetup -o $((start*512)) --sizelimit $(((end-start+1)*512)) /dev/loop1 /mountpoint/file
3. Inspect the RAID member
mdadm --examine /dev/loop1
Additional commands
- Detach a loop device:
losetup -d /dev/loop1
- Assemble the array:
mdadm --assemble --force /dev/md0 /dev/device...
_Originally published in March 2013._