Thanks, gogalthorp, for your meaning.
I know that the snapper is not a backuping tool. But making it one is as simple as
btrfs subvolume create /backup/.snapshots
mkdir /backup/.snapshots/1
btrfs send /btrfs/.snapshots/1/snapshot | btrfs receive /backup/.snapshots/1
and then
mkdir /backup/.snapshots/2
btrfs send -p /btrfs/.snapshots/1/snapshot /btrfs/.snapshots/2/snapshot | btrfs receive /backup/.snapshots/2
(if /backup is mounted external drive formatted with btrfs)
And if I really need to, then rsync to somewhere else.
By the way what do I need the btrfs for? For all the headaches with virtualBox disk images, mysql and sqlite database files? For all the confusion with df output? For all the complexity of /proc/mounts and so on? For time machine, that doesn’t work, if my hard drive loses some sectors?
I have a huge home with a big amount of the stuff, which doesn’t change at all. As I started to use Unison for files synchronisation, it tooks over hour for synchronisation because of files amount (all that uni stuff - IT faculty).
At the end I’ve made a 5 GB tar.gz, which I’m currently not able to open (am too impatient to wait for ark to open it). But even after all of that, all the fotos and Co. the tools like rsync need some time for comparison. (I’m currently using ntfs on external drive).
Reading about “brand new” btrfs I thought: Hey! that’s it! Backup much more efficiently, then rsync ever could! Compression support! (And so much to be learnt to not to rustlol! !)