I'm running ddrescue on a 500 GB drive in an external enclosure. I was copying the image to another external drive, and also using the log file feature.

About halfway through the process, a power blip took the system out. As a result, when I restarted the system, the cloned output file was not accessible ("ddrescue: can't open output file: input/output error"). The external drive I was putting the file on is ntfs, and fsck was not working on it. I put it back onto a windows box, and ran ntfs tools on it. I think this killed my output file, as it showed as being 0 bytes in length when I hooked it back up.

I restarted ddrescue, which picked up where it left off due to the log file. Now my output file says it is 293886755840 bytes and rising.

So, my question is, does ddrescue scan for blocks first, reserving space in the output file and moving it at the end, or does it do it as it goes along, and my image file is toast?

Feel free to correct me, but I'm pretty sure the image is toasted. The first 4k is nothing but nulls, while there is random data stored after the offset approximating the beginning of the ddrescue restart.