My ext4 disk crashed and I'm using ddrescue to try to rescue it. However the crashed disk is 3tb, and I only have another 3tb drive for the image file. I assumed that this image file will be too big for the brand new empty disk, so after 2700gb had been rescued I interrupted the process to try to write the rest of the image to a second disk.

I started the process with

ddrescue -v --no-split /dev/sdc /mnt/red/imagefile /mnt/500gb/logfile

/dev/sdc is the crashed disk /mnt/red is the new hard disk, imagefile is my image /mnt/500gb/logfile is the logfile

Everything looked normal until I interrupted the imaging, so I attempted to continue the image on the second disk using:

ddrescue -v --no-split /dev/sdc /mnt/500gb/imagefile2 /mnt/500gb/logfile

ie, exactly the same command as before, only directing the image file to another disk

However, after just a few seconds ddrescue quits and complains that there is no space left on the device, but the device is empty apart from the logfile. What I don't understand is that imagefile2 is created immediately with a size of 2,732,050,104,320 (assume bytes), which is roughly the same size as the amount rescued in the first image file.

I cannot find any instructions anywhere as to why this might be. I assumed the second image file would be roughly the same size as the remaining amount to be rescued. ie 250gb.

  • 2
    Why not just use the entire disk, rather than an image file? Also, don't use FAT/NTFS (lack of sparse file support). – frostschutz Feb 23 '15 at 11:08
  • Are there any advantages of cloning the disk over creating an image file? Which is better? This 3tb of data is very important to me. As for sparse file support, that is indeed the problem :) – Mike Timberlake Feb 23 '15 at 11:15
  • 1
    ddrescue knows it's dealing with a broken image and tries very hard to recover as much data as possible. IME some cloning programs assume a clean source disk and don't handle bad reads as well as they might. As suggested by frostschutz, use the disk as your image target for ddrescue and write the logfile somewhere else (maybe /tmp on your system rescue's ramdisk). – roaima Feb 23 '15 at 15:34
  • thanks! that's what i decided to do, after discovering what a PITA it is to have a split image file. – Mike Timberlake Feb 23 '15 at 20:27
  • One last quick question. I successfully cloned the entire drive to the new drive, and ddrescue reported only about 20 errors, total size a few megabytes. I started a second pass to retry the failed sectors, but incorrectly specified the log file. I stopped the scan after a few minutes. Can I safely assume that I haven't ruined my clone, even through the scan was restarted? Or should I lose another 30 hours recloning the whole thing again..? – Mike Timberlake Feb 25 '15 at 10:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.