brainbitsgmbh/mysql-backup repository overview

⁠MySQL Backup

This image allows a simple backup strategy for mysql and mariadb databases. Backups are created via mysqldump and gziped, and can be restored with this image by just changing a single environment variable.

⁠Backup strategy

This image follows a simple strategy for keeping or removing backuped dumps:

• It checks whether there already is a file in the output directory that was created on the current day.
  • If no, a new backup will be created and treated (and named) as a "daily" backup.
  • If yes, a new backup will be created and treated (and named) as an "hourly" backup.
• Depending on whether it created a "daily" or "hourly" backup, it might remove certain older backup files:
  • Daily: It will remove all daily backups (files starting with the name daily-*) that are older than a specified amount of days.
  • Hourly: It will remove all hourly backups (files starting with the name hourly-*) and only keeps a specified amount of files.

This strategy ensures, that there are always a certain amount of backup files available from the last couple of days, but also that there are at least a few recent backups available too, without storing too many backups.

⁠Variables

The following environment variables can be used to configure the script:

⁠Examples

Example creating and restoring a backup using docker run:

# Create a backup. It will automatically create a link called `latest` to that backup.
docker run -it --rm -e "MODE=backup" -e "DATABASE_DSN=mysql://MyDbUser:[email protected]/my-database" -v "$(pwd)/backup:/backup" brainbitsgmbh/mysql-backup:latest

# Restore the `latest` backup.
docker run -it --rm -e "MODE=restore" -e "DATABASE_DSN=mysql://MyDbUser:[email protected]/my-database" -v "$(pwd)/backup:/backup" brainbitsgmbh/mysql-backup:latest

⁠Troubleshooting

⁠The script cannot connect to my database

Make sure that you enter a valid DSN that can also be reached from withing the docker container the script is running in. If your DSN includes localhost it will likely not work, because the container cannot reach the database on your local machine via localhost. In that case you can try host.docker.internal instead, as seen in the example above.