How incremental backups work
The first restic backup copies all your data. Later backups are much faster because restic copies only what changed. This topic explains how.
Chunks and deduplication
Restic splits each file into variable-size chunks and gives every chunk an identifier based on its content. Before it stores a chunk, restic checks whether the repository already has a chunk with the same identifier. If it does, restic skips the copy. This process is called deduplication.
What this means for each backup
- The first backup stores every chunk.
- A later backup reads your files again but stores only new chunks.
- A renamed or moved file adds almost no data, because its chunks already exist.
Every snapshot is complete
Although restic stores only changed data, each snapshot describes the full set of files. To restore, you pick one snapshot. You never need to combine a full backup with a chain of smaller ones.