Backup Targets and Storage

Step by step how-to instructions on installing and running KoBo Toolbox on a standalone server.
Jump to: navigation, search


In general, web applications keep their data either within services (such as RDBMSes or a Key-Value Stores), or in the filesystem, using the Host's folder/file structure).These are, of course, our backup targets: the databases and the files..

In a more traditional application setup, backing up the logic (i.e. the application code) would also be necessary. Since Kobo Toolbox runs off Docker images, the application logic (at any particular point in time) is available in the Repository, so no code backups are required.

In the Kobo-Docker setup, three backup targets are identified:

No Description Container Location Export format
1 User and system generated files kobodocker_kobocat_1 /srv/src/kobocat/media Folders and files in .tar.gz archive
2 Postgres database kobodocker_postgres_1 postgress service .pgdump binary
3 Mongo database kobodocker_mongo_1 mongo service .json and .bson in tar.gz archive

Target 1 is also available in $KOBO_DOCKER/.vols/kobocat_media_uploads/ on the Host. It is the same folder as the one inside kobodocker_kobocat_1, and it is a matter of convenience to decide which one to access for backup.

Target 3 is dispensible, redundant data that can be recreated from the rest of the system state; it is backed up primarily to faciliate swift recovery.

For detailed review of backup functions, please refer to Backup Procedures.


For this particular instance, backup storage was allocated on a Windows server on the local network. The backup storage is getting its own regular backup, per MSF ICT rules and regulations.

Backup storage is published as a Windows share on // This share is mounted on /media/kobo-backups/ on the Host.

To permanently mount backup storage onto KoBo server's filesystem, follow the steps laid out in Backup setup.