I'm finding ddrescue a bit lacking for my particular use case scenario, one which I find cannot be uncommon:
I'm working with some sensitive data on a crashed hard drive (piece of shit WD Green drive) and want to fill the drive with random data before sending it back to the manufacturer for warranty replacement. Preferably I want to do several passes to make sure data recovery is (sufficiently) impossible. I have already copied everything I can to a new hard drive and simply want to fill all writable sectors using with random data. I'm using the following command line:
ddrescue --verbose --force --fill=+ /dev/urandom /dev/sdb recovery.log
Where recovery.log is the log from my recovery operation. However, ddrescue constantly quits because of write errors:
ddrescue: write error: Input/output error
This is quite irritating because I constantly must go in to the log file and manually change the current position to make ddrescue skip the faulty sector (which obviously wasn't faulty when I did the original rescue). It would be so much simpler if I could give ddrescue a command-line option to skip ahead on write errors in fill mode, just as it does in recovery mode on read errors.
----
JayC