CephFS snapshots

Use CephFS snapshots to create an immutable, point-in-time view of the file system. CephFS directory. CephFS snapshots are asynchronous, and you can choose which directory snapshots are created in.

Create snapshots in one of the following ways:

With the Ceph Dashboard.
For more information, see Managing CephFS snapshots.
By invoking the mkdir command within the .snap hidden directory
By using CLI commands.

Snapshots can also be automatically created at specific intervals through snapshot schedules and policies.

Before creating CephFS snapshots, be sure to have CephFS deployed and root-level access to the Ceph MetaData Server (MDS) node.

The Ceph File System (CephFS) snapshotting feature is enabled by default on new Ceph File Systems, but must be manually enabled on existing Ceph File Systems. CephFS snapshots are asynchronous and are kept in a special hidden directory in the CephFS directory named .snap. You can specify snapshot creation for any directory within a Ceph File System. When specifying a directory, the snapshot also includes all the subdirectories within it.

Warning: Each MDS cluster allocates the snap identifiers independently. Using snapshots for multiple Ceph File Systems that are sharing a single pool causes snapshot collisions, and results in missing file data.

For more information, see Deploying the CephFS and CephFS snapshot schedules.

CephFS snapshots usage

Use CephFS snapshots for data protection and disaster recovery.

Data protection: Snapshots allow users to recover files from specific points in time. It is useful for creating backups without disrupting ongoing operations.
Disaster recovery: Snapshots are used for CephFS mirroring which enables to replicate CephFS data asynchronously to another Ceph cluster in another location.

CephFS limitations

Before using CephFS, get to know the CephFS limitations.

Snapshot names: Snapshot names cannot start with an underscore (_) and cannot exceed 240 characters.
Performance impact: While creating snapshots is fast, the performance impact can be noticeable during heavy I/O operations due to the asynchronous nature of data flushing.