This page documents in a very simplified way the concepts used by plakar.
At the core of plakar are repositories.
A repository is a storage unit within which backups are stored. You can have one or many different repositories, some encrypted and others not, each backed by your local filesystem, a database or even a remote server.
They should be considered as locations, rather than directories: you point the utility there so it can store snapshots or retrieve data from snapshots.
The following command will for example create a repository in /tmp/plakar:
$ plakar on /tmp/plakar create -no-encryption
$
While the following command will list the snapshots that are part of that repository:
$ plakar on /tmp/foobar ls
$
None here as the repository was just created and is empty.
Repositories are useless without backups.
The plakar utility creates backups from directories.
It scans their content and builds an internal representation of what’s necessary to rebuild an exact copy,
both of the filesystem structure and files,
then records that representation as a versioned snapshot in a repository:
$ plakar on /tmp/foobar push /bin
$ plakar on /tmp/foobar ls
2022-04-08T21:38:50Z 21a403b1-b022-40dd-95ad-5b78a90360dd 13 MB /bin
$
Snapshots are space efficient as they may contain the same files as other snapshots without having to store their own copies, creating multiple snapshots of the same directory will consume roughly the size a single copy would consume with a little overhead due to snapshot-specific metadata:
$ plakar on /tmp/foobar push /bin
$ plakar on /tmp/foobar push /bin
$ plakar on /tmp/foobar push /bin
$ plakar on /tmp/foobar push /bin
$ plakar on /tmp/foobar ls
2022-04-08T21:38:50Z 21a403b1-b022-40dd-95ad-5b78a90360dd 13 MB /bin
2022-04-08T21:38:51Z 976ed175-206a-418d-b960-7d0736d41b46 13 MB /bin
2022-04-08T21:38:51Z 8271734e-882c-47b4-812a-5bef3455281c 13 MB /bin
2022-04-08T21:38:52Z 273c9764-eae3-4e4f-8bcd-39538b684b49 13 MB /bin
2022-04-08T21:38:52Z 70e1de7a-e410-4ad4-b441-f4f40977232b 13 MB /bin
$ du -sh /tmp/foobar
4.0M /tmp/foobar
$
In some situations, such as above, the space consumed will even be smaller than that of a single snapshot as the snapshot may contain duplicate informations itself, or may compress well.
Contrarily to incremental backups, snapshots are truly independant one from another and do not assume the existence of any other snapshot. It is possible to compare the differences between a snapshot and any other snapshot, just as it is possible to delete a snapshot without affecting any of the subsequent snapshots:
$ plakar on /tmp/foobar rm 21a403b1 976ed175 8271734e 273c9764
$ plakar ls
2022-04-08T21:38:52Z 70e1de7a-e410-4ad4-b441-f4f40977232b 13 MB /bin
$ plakar on /tmp/foobar pull 70e1de7a:/bin
$ ls bin
[ csh echo ksh mkdir rm sync zsh
bash dash ed launchctl mv rmdir tcsh
cat date expr link pax sh test
chmod dd hostname ln ps sleep unlink
cp df kill ls pwd stty wait4path
$