Disk copy of failing drive with Fedora 14

4

I have a WD disk drive of 1 Tera Byte capacity with Fedora 14 installed. On it I also have Virtual Box with a Windows XP guest installed.

Now BIOS startup and Fedora is reporting that disk failure is imminent, so I bought another 1 TB hard drive to copy the whole drive.

Now, how can I transfer the whole image to this other hard disk of similar capacity? Both are 1 TB, the original one is failing so instead of fresh install, update and lot more and then finally copy data... somehow I want to image the first hard disk completely to the new drive so that I can wipe out the defective hard drive before sending it away for replacement.

  • I tried Norton Ghost 8.3 for disk-to-disk copy, but that failed.
  • I tried CloneZilla, with normal, easy user mode with disk-to-disk copy, and that copied the first boot partition (500MB) within a minute despite reporting 12+ hours (as reported in progress bar) to copy 1TB to other drive, but after 7-8 hours it hangs (keyboard Caps Lock and Scroll Lock keys started blinking) and brings up a non-responsive command prompt.

...I am assuming because of failing drive, some part may be corrupt and non-readable, and that's probably where it hangs.

I wondering how I can clone the disk drive? Any help?

cd10anil

Posted 2011-05-10T03:47:48.430

Reputation: 41

Answers

3

Get dd_rescue which can create a 1:1 clone and which doesn't stop for failing blocks. Be ready for some longer waiting times, though. dd_rescue will try to read each block several times. On top of that, the OS will retry failed blocks as well. So it will hang a long time for each bad block but it will eventually pass over them.

Note that you should not delete the images you already have! Keep them because the drive probably fails more and more, as you try to create the image, so you may be able to save some broken files from the images you have.

If you're impatient, you can try dd_rhelp. When it encounters a bad sector, it will skip ahead a couple of megabytes and try to read the data after the bad area. When it reaches the end of the device, it will go back, trying to read more of the area between the bad sectors and the end of the disk.

That way, it will quickly get most of the good sectors and get ever slower as it approaches the bad sectors.

Aaron Digulla

Posted 2011-05-10T03:47:48.430

Reputation: 6 035

0

You can try Drive Snapshot, which can perform a raw copy. I've used it to transfer data off of failing disks for clients over the years, and the one feature I really REALLY like is that it is smart about dealing with bad sectors -- it continues after the failure and tracks the sector number in a plain/text log file.

  Drive Snapshot Disk Imaging
  http://www.drivesnapshot.de/

If your file system isn't supported, at least you can get a raw copy of the entire drive. The catch will likely be that you'll need to restore to a drive of the same model unless you can find one that has the same specifications (e.g., cylinders, heads, sectors per track, etc.). Experimentation with your new drive will help greatly in determining if this will work for you.

Do keep in mind, however, that every time you access a failing disk, you are probably reducing the chances of successfully recovering your data. The very best approach is to always work with a copy (which is what you're obviously trying to accomplish).

Randolf Richardson

Posted 2011-05-10T03:47:48.430

Reputation: 14 002