Tool to make remote-incremental backups which consist of a mirror and historical snapshots of the target data.
These snapshots are created using hard links and can be taken as often as needed as they are cheap both in time and
space. rincr can manage these snapshots, clean up unnecesary data and also restore files to earlier versions.
Note: This project was previously called kbackup. See this issue for more information.
rincr supports a similar interface to rsync:
rincr [[USER@]HOST:]SRC... [[USER@]HOST:]DEST
When this command is run rincr checks to see if there are existing backups at the destination. If there are, a copy of
the latest backup is created using hard links. Then new changes (if any) are synced from the source.
Incremental backups are kept at DEST/SRC/.rincr. They are fully browsable and take no extra space for files that have
not changed between versions. Each file acts as a full-backup and can be copied back out manually to restore data to an
older version.
Download the right binary for your OS and Architecture (all binaries available on Github):
Make sure it is executable and then place it anywhere in your $PATH:
chmod +x rincr
mv rincr ~/bin/rincr # Use whatever location you put your user binaries in.
Test that it works:
rincr --version
Note: This software is not yet stable; there may be backwards-incompatible changes before v1. Use at your own risk.
Create an incremental backup of ~/mydata at the remote location myserver:~/backups/mydata:
rincr ~/mydata myserver:~/backups
Create an incremental backup of server1:~/mydata & server2:~/otherdata at the local locations ~/backups/mydata &
~/backups/otherdata respectively:
rincr server1:~/mydata server2:~/otherdata ~/backups
If we want to clean up older backups, pass the --prune option:
rincr --prune ~/mydata myserver:~/backups
This will apply the default retention rules and delete extra backups. Pruned backups will be printed out. The default retention rules are:
If you just want to prune backups without syncing new data, you can use:
rincr prune myserver:~/backups/mydata
Note: When pruning data you have to provide the path to the actual backup destination which includes the target name
mydata.
If you want to override the default retention rules, you can use the options:
rincr --prune --hourly=12 --daily=15 --monthly=6 --yearly=5 ~/mydata myserver:~/backups
To restore files from a backed-up repository, we can use:
rincr restore --latest myserver:backups/mydata path-to-restore output-path
rincr will check backups from latest to oldest until it finds a matching path. It will then copy that path recursively
into the output location. We can restore single files or full directory trees this way. Since rsync is used for the
the underlying transfer, only necessary files are transferred. Mutiple paths can be restored in one go:
rincr restore --latest myserver:backups/mydata file-1 file-2 output-path
Any paths that are not found will skipped. We can also configure how old of a backup we want to fetch:
rincr restore --from=12h myserver:backups/mydata path-to-restore output-path
It will find the closest backup to the given duration (from current time) and copy the path if it finds one.
$ ls -A ~/mydata
README.md new-file.txt
$ ssh desktop-1 ls ~/Backups/mydata
README.md .rincr
$ ssh desktop-1 ls ~/Backups/mydata/.rincr
2024-01-12T09-12-32 last
$ rincr ~/mydata desktop-1:~/Backups
...
$ ssh desktop-1 ls ~/Backups/mydata
README.md new-file.txt .rincr
$ ssh desktop-1 ls ~/Backups/mydata/.rincr
2024-01-12T09-12-32 2024-01-23T18-26-10 last
# File encryption - supports deduplication but less secure
rincr --encrypt ~/mydata myserver:~/backups
# Folder encryption - no deduplication but completely secure
rincr --encrypt-folder ~/mydata myserver:~/backups