a tale of fail and win (image recovery/management under linux)

  1. use git-annex-assistant to create backups on several destinations
    1. use test-repo first, do some tests
    2. try on smaller directories w/ actually valuable data, create backups first
    3. annex-ize several other directories
    4. remove picture-backup from external HDD to make space for new backup via git annex (very bad ida)
    5. annex-ize several GB of pictures dating back to 2004 (RAWs and JPGs)
    6. fail somehow several times, remove .git directory start anew
    7. (do some other stuff)
    8. get back to the picture-dir, realize that it is empty (besides some folders), .git directory contains nothing
  2. use ntfsundelete, and some proprietary tools to recover (only marked as) deleted files from the ntfs volume (900 GB)
    • use git annex fsck on the recovered .git data, get only some pictures back, not very much (about 2k files)
  3. use photorec on several runs to recover .jpg and .cr2 (RAW) data
  4. try to use picasa on the files to get some sorting (and kick out unwanted data as images from games etc.)
    • picasa somehow mangles the raw-files :(
    • picasa does not properly use the exif-provided file-creation date, but a mixture of that and the files’ date w(
  5. fiddle around with exiftool to get back the timestamp from the files’ exif-data
    find . -type f -name "*.jpg" -exec exiftool  -FileModifyDate\<DateTimeOriginal {} \;
  6. try digikam
    1. somehow works
    2. slow on previews when using ‘import from files’
    3. slow on DB handling
    4. hangs itself when moving about 6k (?) files from one folder to another
    5. switch to MySQL as backend
      • somehow fail, try google
      • realize that the internal MySQL server won’t do, install external one
      • use ‘settings’->’Database migration’ before switching via the config
    6. speed is better
    7. use the duplicate detection to remove redundant files (takes time …)