Benji, backup me!¶
Benji Backup is a block based deduplicating backup software. It builds on the excellent foundations and concepts of backy² by Daniel Kraft.
While Benji can backup any block device or image file (this includes LVM logical volumes and snapshots) it excels at backing up Ceph RBD images and it also includes preliminary support to backup iSCSI targets.
Benji is written in Python and is available in PyPI
for installation with
pip. Benji also features a generic container image with all
dependencies included as well as an image and Helm chart to integrate Benji into a
Kubernetes environment to backup Ceph RBD based persistent
The documentation is available here.
Benji is slowly nearing beta quality. It passes all included tests. The documentation isn’t completely up-to-date. Please open an issue on GitHub if you have a usage question that is not or incorrectly covered by the documentation. And have a look at the CHANGES file for any upgrade notes.
Benji requires Python 3.6.5 or newer because older Python versions
have some shortcomings in the
concurrent.futures implementation which lead to an
excessive memory usage.
The Kubernetes integration is currently in the process of being completely rewritten
to use an operator based approach. In the meantime the
image together with the Helm chart already provides a solid way of backing up
persistent volumes provided by Ceph RBD. Benji will detect both normal RBD
volumes and volumes provisioned by Rook’s FlexVolume provisioner.
- Small backups
Benji deduplicates all data read and each unique block is only written to the storage location once. The deduplication takes into account all historic data present on the backup storage and so spans all backups and all backup sources.
In addition Benji supports fast state-of-the-art compression to further reduce the storage space requirements.
- Fast backups
With the help of snapshots and the
rbd diffcommand Benji only backups blocks that have changed since the last backup when used with Ceph RBD images. The same mechanism can be extended to other backup sources.
- Fast restores
Sparse blocks are be skipped on restore providing fast restores of sparsely populated disk images.
- Low bandwidth requirements
As only changed and not yet known blocks are written to the backup storage, the bandwidth requirements for the network connection between Benji and the storage location are usually low. Even with newly created block devices the traffic to the backup storage location is generally small as these devices mostly contain sparse blocks. Enabling compression further reduces the bandwidth requirements.
- Support for a variety of backup storage locations
Benji supports AWS S3 as a backup storage location and it has options to enable compatibility with other S3 implementations like Google Storage, Ceph’s RADOS Gateway or Minio.
Benji also supports Backblaze’s B2 Cloud Storage which opens up a very cost effective way to store backups.
Benji is able to use any file based storage including external hard drives and network based storage solutions like NFS, SMB or even CephFS.
Multiple different storage locations can be used simultaneously and in parallel to accomodate different backup strategies.
Benji supports AES-256 in GCM mode to encrypt all data blocks on the backup storage. By using envelope encryption every block is encrypted with its own unique random key. This makes plaintext attacks even more difficult.
Each data block in Benji is protected by a checksum. This checksum is not only used for deduplication but also to ensure the integrity of the whole backup. Long-term availability of backups is ensured by regularly checking existing backups for bit rot.
- Integrated NBD server
Benji brings its own NBD (network block device) server which makes backup images directly accessible as a block device - even over the network. The block device can be mounted if it contains a filesystem and any individual files needed can be easily restored even though Benji is a block based backup solution.
Benji can also provide a writable version of a backup via NBD. This enables repair operations like
fsck. The original backup is not changed in this case. All changes are transparently written to a new backup via copy-on-write and this new backup can be restored just like any other backup after the repair is complete.
Benji supports running multiple operations simultaneously. Instances can be distributed across different hosts or containers without the need for a central server.
Benji comes with a module framework to easily add new protocols for accessing backup sources or storages. New compression and encryption algorithms are also easily integrated into Benji.
- Quick Start
- Retention Policy Enforcement
- Containerized Benji
- Filter Expressions
- Data Layout
- For Backy² Users