Rsync help

Question

Alright SoF, this is a little hard to explain, but i'll try my best.

I have 2 identical drives. Let's call "S:" for source and "D:" for destination

S: is the drive i keep all my files in (images, music, videos, documents, etc), and D: is a backup HD i (manually) back up every Sunday night.

What i'd like to do is, backup S: into D:, with a few rules.

Like, i said, i do backups once a week. This means that throughout the week, files get added, deleted and moved around from a folder to another.

1. Only copy new files, or files that have been modified(would ned to check file's metada)

2. At the end of the back up, D: would have to end up being identical to S:.

Meaning, if i moved a file from folder "A" to folder "B" in S:, the back would see that the file is no longer in folder "A", and would have to delete it, to make that folder identical to S:.

step 2 was probably poorly explained., so here's a better explanation. This is how I plan on doing things if rsync can't do it.

In python, I would create a script that does the following (in order):

1. Compares D: to S: - The script would first traverse D:. each time it enters a directory, it looks at that same directory in S:. It then looks at the files. If a file is in D: but not in S:, that means the file has been deleted or renamed or moved around in S:. Therefore, delete that file from D: (repeat this process for all folders)

2. Now that D: have the exact same files (or less if they were deleted in steps above), start copying. First check if the current file in S: exists in D:, if not, then copy. If it does, check metadata. if it has been modified, copy and overwrite.

Sorry if I didn't explain this too well, it's 4:15am, and frankly, im tired of manually deleting all of D: and copying files over from S: =/

Answer 1

I think you're looking for the --archive switch along with --delete to rsync. That is, if I'm right in assuming that your description means "I would like D: to have the same contents as S: after running a backup, without having to delete everything on D: first".

I'm also assuming that although you're using Windows notation for drive letters, that you're not actually using Windows.

Answer 2

Here's a script I wrote to backup my linux machine to a USB drive.

#!/bin/sh

rsync -a \
  --progress \
  --hard-links \
  --whole-file \
  --delete \
  --delete-after \
  --delete-excluded \
  --stats \
  --filter='- *.log' \
  --filter='- /dev' \
  --filter='- /boot' \
  --filter='- /media/' \
  --filter='- /mnt' \
  --filter='- /net' \
  --filter='- /proc' \
  --filter='- /tmp/' \
  --filter='- /var/log/' \
  / /media/disk/middle-earth

The --filter lines exclude files/subdirectories that I don't want to sync.

You can use this as a starting point to craft your own.