cp2dd

Why?

This script was created for the following case:

• You had an external HD you were copying files from, and these files failed occasionally.
• You don't want to ddrescue the entire disk, because: a) you might have enough space to do it, or b) you won't be able to tell which files were sitting around the bad blocks from the ones that are healthy.
• I created this script because I needed it (the exact above scenario).

How?

1. After you run your cp -R, you will get a list of files that failed to copy. Please save this list as input.txt. E.g.:

cp: /Volumes/SAMSUNG/D/Backup/Alisson/Desktop/Organizar/FOTOS/DSC00329.JPG: Input/output error
cp: /Volumes/SAMSUNG/D/Backup/Alisson/Desktop/ultimas/DSC03950.JPG: Input/output error

2. Install dd (probably built-in) and ddrescue. You might want to use homebrew for that (http://brew.sh/).

3. Install Python3 (script wasn't tested on earlier versions).

4. Run cp2dd.py using the following parameters:

#use dd to copy
./cp2dd.py -i input.txt -o output.sh -f /prefix/path/from/ -t "/prefix path/to/"

#use ddrescue to copy
./cp2dd.py -i input.txt -o output.sh -f /prefix/path/from/ -t "/prefix path/to/" -udr

#sample help output
$ ./cp2dd.py -h
usage: cp2dd.py [-h] -i INPUT [-o OUTPUT] [-f FROM_DIR] [-t TO_DIR] [-la]
                [-lf] [-ld] [-udr]

Quickie to transform cp -R errors in dd commands for further desperate
attempts of recovery.Please use double quotes to enclosure paths with spaces.

optional arguments:
  -h, --help            show this help message and exit
  -i INPUT, --input INPUT
                        Input file name
  -o OUTPUT, --output OUTPUT
                        Output file name
  -f FROM_DIR, --from_dir FROM_DIR
                        From prefix path
  -t TO_DIR, --to_dir TO_DIR
                        To prefix path
  -la, --listall        List all files and directories after processing. Will
                        not write to output.
  -lf, --listfiles      List all files after processing. Will not write to
                        output.
  -ld, --listdirectories
                        List all directories after processing. Will not write
                        to output.
  -udr, --useddrescue   Use ddrescue instead of dd.

4. run the resulting output.sh, this will attempt to copy the failed files again one by one, and try to recover data (when using ddrescue).

Features:

• Utilises the recommended conv=noerror,sync for dd copy.
• When using ddrescue, tries to read 6 times form the source (3 forwards and 3 reverse).
• When using ddrescue, utilises a logfile to save data already copied accross 3 runs as to save time.
• Allows you to list folders and filers (-la, -lf or -ld). When listing no attempt to copy is made.
• This script does not attempt to copy folders accross, only files. It tries to detect files by the extension (e.g.: .bmp is considred a file). I do not try to read file information to do so in an attempt to touch the origin disk the less as possible.

Disclaimer:

I recommend you to read the code and understand what the generated bash script does. Recovering data is hard and I cannot say this script is gonna help every time. If your data is important, do not attempt to rescue directly as you might damage the drive even further. Tested on a MacOS X.