Puedes leer este artículo en castellano aquí

Why?

Because who doesn't need a database backup? ;)

Background

For my personal project, FacturApp, I use PostgreSQL because it's awesome and also because is free. From time to time I need to restore the prod db to my local instance to have true data and make tests more reliable.

Because I love automation, I decided to have a daily backup just in case and when I need to have the current data, I would just download the file from S3.

How?

Steps

The TLDR part would be

1. Find the container's ID
2. Save current time in a variable called now
3. Save the db backup using the now variable for the name
4. Upload backup to S3
5. Locally delete files older than 30 days

The Code

#!/bin/bash
docker_id="$(docker ps -aqf 'name=postgres')"
echo "${docker_id}"
now=`date +%Y%m%d%H%M%S`
docker exec ${docker_id} /usr/local/bin/pg_dump -U theuser -Fc thedb > /root/backups/facturapp_${now}.backup
# Upload file to S3
/usr/local/bin/aws s3 cp /root/db_daily/facturapp_${now}.custom s3://db-backup-bucket/facturapp_${now}.custom --sse AES256
# Deleting old files
find /root/db_daily -type f -mtime +30 -exec rm {} \;

Automate it!

The final step would be to create a cron job by running crontab -e and adding something like
0 0 * * * /root/dailybackup >> /root/logs/backup.log
to have to script running each day at midnight.

And that's how I save a backup to an S3 bucket every day at midnight :)

The /root/logs/backup.log part is just in case you want to save the log for each time the script runs.