[rdiff-backup-users] Incremental Restore

[Kosta Welke]

Hi!

I'm currently trying to move from a "incremental rdiff-backup on same 
filesystem" to "modern filesystem with snapshots". (Yes, I know, this isn't a 
real backup - it's not offsite yadda yadda. This is protection against "Wife 
deletes her thesis by accident", no against "House burns down". Against that, 
there's also offsite backups - but that's not the topic here.)

I plan to do the following:

1. Restore full backup of oldest existing version from $backup to $new
2. Take a snapshot of $new
2. For each increment:
2.1. List all changed/added files using:

zgrep -e '1\( [NA0-9]\+\)\{3\}$' 
$backup/rdiff-backup-data/file_statistics.$timestamp.data.gz

2.2. Feed that into rdiff-backup using --include-filelist-stdin and --exclude * 
and --force as files might already exist. (Of course, some awk magic is needed 
here)
2.3. Take a snapshot of $new

Do you think this is a feasible approach? If not, what would you suggest for an 
"incremental restore"?

Cheers,
Kosta