Incremental Backups

Save time and storage by backing up only what changed.

An incremental backup stores only the files that changed since the previous run, instead of re-uploading everything each time. For large file sets that change slowly, this cuts both backup duration and storage cost substantially.

How it works

SimpleBackups uses GNU tar with a listed-incremental snapshot file kept on your server:

1. The first run is a full backup — every file and folder, and a snapshot file recording their state.
2. Subsequent runs compare against that snapshot file and archive only what changed. Each run produces one compressed tar file containing just the changes, uploaded to your storage as usual.
3. Periodically a new full backup is taken, starting a fresh chain.

Each run is tagged with an increment ID whose final segment is the level — .0 marks a full backup, and higher numbers are the increments built on top of it. You can see these in the backup's Logs tab.

!Restoring needs the whole chain

To restore to a given point in time you need the full backup (.0) and every increment up to that point. A single incremental file is not restorable on its own.

This has a direct consequence for retention: retention values are normalised to account for the chain, but you should never manually delete files from an incremental chain in your storage. Removing a full backup makes every increment after it unusable.

When a new full backup is taken

A new full backup starts automatically when there is no existing chain, or once the number of increments reaches the schedule's retention limit. This keeps chains bounded, so a restore never has to replay an unlimited number of increments.

Requirements and limits

• Plan — incremental backups require the Lite plan or higher. See Subscriptions.
• One schedule per backup — an incremental backup can have only one schedule, on every plan, because the increment chain depends on a single ordered sequence of runs. Multi-scheduling is not available for incremental backups.
• GNU tar on restore — the restore procedure requires GNU tar on the target machine. BSD/macOS tar cannot read listed-incremental archives, and the restore script will stop rather than produce a partial result.

+When not to use incremental

Incremental backups trade restore simplicity for speed and storage. If your data set is small enough that a full backup is cheap, or you want every run to be independently restorable, stick with full backups. Incremental pays off on large, slow-changing file sets.

Restore a file or server backupDownload and extract a file backup, including incremental chains.›RetentionHow retention interacts with incremental chains.›