django-caretaker 0.1.4

A Django app for backups using S3 storage.

django-caretaker

django-caretaker ('The Caretaker') is a Django app that backs up your database and media files to a versioned AWS S3 bucket. It comes with the Terraform files to provision the S3 bucket and provides management commands to schedule regular backups.

Install

To install the module, use pip:

pip install django-caretaker

Add 'caretaker' to your installed apps in your Django settings file.

Add 'path('caretaker/', include('caretaker.urls')),' to your urls.py file to enable the /caretaker/list view.

Setup and Configuration

Configure an AWS S3 Bucket and IAM User for Backup

Next, ensure that you have a working AWS cli client and configure it if not.

Set the BACKUP_BUCKET variable in your settings.py file. This must be a globally unique name for the S3 bucket. You should also set the MEDIA_ROOT folder so that we know what to back up:

BACKUP_BUCKET = 'caretakertestbackup'
MEDIA_ROOT = '/var/www/media'

Generate and run Terraform configuration in your home directory:

./manage.py get_terraform --output-directory=~/terraform_configuration
cd ~/terraform_configuration
terraform init
terraform apply
terraform output --json

Note down the AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY and put them in your settings.py file:

AWS_ACCESS_KEY_ID = 'PUT_ACCESS_KEY_HERE'
AWS_SECRET_ACCESS_KEY = 'PUT_SECRET_ACCESS_KEY_HERE'

Install the Backup Script in Cron

To install a cron line that will run the backup daily at 15 minutes past midnight on the server, run:

./manage.py install_cron --action=test
./manage.py install_cron

Usage

Caretaker provides a number of management commands that can be accessed from manage.py:

Run Backup

This is the most important command. It backs up your database and your media files to the remote store.

usage: manage.py run_backup [-h] [--output-directory OUTPUT_DIRECTORY] [--version] [-v {0,1,2,3}] [--settings SETTINGS] [--pythonpath PYTHONPATH] [--traceback] [--no-color] [--force-color] [--skip-checks]

Creates a backup set and pushes it to S3

Example usage:

manage.py run_backup --output-directory=~/backup

Push Backup

This command pushes a backup to the server.

usage: manage.py push_backup [-h] [--backup-local-file BACKUP_LOCAL_FILE] [--remote-key REMOTE_KEY] [--version] [-v {0,1,2,3}] [--settings SETTINGS] [--pythonpath PYTHONPATH] [--traceback] [--no-color] [--force-color]
                             [--skip-checks]

Pushes the backup SQL to the S3 store

Example usage:

manage.py push_backup --backup-local-file=/home/obc/backups/data.json --remote-key=data.json

Pull Backup

This command retrieves a backup file from the server. You must also specify the version you wish to retrieve.

usage: manage.py pull_backup [-h] [--backup-version BACKUP_VERSION] [--out-file OUT_FILE] [--remote-key REMOTE_KEY] [--version] [-v {0,1,2,3}] [--settings SETTINGS] [--pythonpath PYTHONPATH] [--traceback] [--no-color]
                             [--force-color] [--skip-checks]

Pulls a specific backup SQL from the S3 store

Example:

manage.py pull_backup --remote-key=data.json --backup-version=jB1dtbf1qraDQhBlKGGDXKAZugEnT2KB --out-file=/home/user/data.json

Restoring a Backup

Restoring a backup consists of the following steps. First, find the backups that you want:

manage.py list_backups --remote-key=data.json
manage.py list_backups --remote-key=backup.tar.gz

You can use grep to find a specific date.

Then pull the files down:

manage.py pull_backup --remote-key=data.json --backup-version=<INSERT_BACKUP_VERSION_ID> --out-file=/home/user/data.json
manage.py pull_backup --remote-key=backup.tar.gz --backup-version=<INSERT_BACKUP_VERSION_ID> --out-file=/home/user/backup.tar.gz

Unzip backup.zip and replace the media folders with the results.

Reload the database:

manage.py loaddata /home/user/data.json

Credits

• AWS CLI for interactions with AWS.
• Django for the ORM and caching system.
• django-dbbackup for hints and tips.
• Git from Linus Torvalds et al.
• .gitignore from Github.
• Rich for beautiful output.
• Terraform by Hashicorp.