dzangolab/postgres-13-backup-s3 repository overview

⁠postgres-13-backup-s3

Backup PostgreSQL 13 to S3 (supports periodic backups)

⁠Usage

Docker:

$ docker run -e S3_ACCESS_KEY_ID=key -e S3_SECRET_ACCESS_KEY=secret -e S3_BUCKET=my-bucket -e S3_PREFIX=backup -e POSTGRES_DATABASES=dbname -e POSTGRES_USER=user -e POSTGRES_PASSWORD=password -e POSTGRES_HOST=localhost dzangolab/postgres-13-backup-s3

Docker Compose:

postgres:
  image: postgres:13
  environment:
    POSTGRES_USER: user
    POSTGRES_PASSWORD: password

pgbackups3:
  image: dzangolab/postgres-13-backup-s3
  links:
    - postgres
  environment:
    SCHEDULE: '@daily'
    S3_REGION: region
    S3_ACCESS_KEY_ID: key
    S3_SECRET_ACCESS_KEY: secret
    S3_BUCKET: my-bucket
    S3_PREFIX: backup
    POSTGRES_DATABASES: dbname
    POSTGRES_USER: user
    POSTGRES_PASSWORD: password
    POSTGRES_EXTRA_OPTS: '--schema=public --blobs'

⁠Environment variables

• ENCRYPTION_PASSWORD if not empty, backups will be encrypted using this password
• POSTGRES_DATABASES list of databases you want to backup. ALL to backup all databases required
• POSTGRES_EXTRA_OPTS pg_dump options (default: '')
• POSTGRES_HOST the postgres host required
• POSTGRES_PORT the postgres port (default: 5432)
• POSTGRES_USER the postgres user required
• POSTGRES_PASSWORD the postgres password required
• POSTGRES_PASSWORD_FILE path to file containing the postgres password; alternative to POSTGRES_PASSWORD
• S3_ACCESS_KEY_ID your AWS access key required
• S3_ACCESS_KEY_ID_FILE path to file containing your AWS access key; alternative to S3_ACCESS_KEY_ID
• S3_SECRET_ACCESS_KEY your AWS secret key required
• S3_SECRET_ACCESS_KEY_FILE path to file containing your AWS secret key; alternative to S3_SECRET_ACCESS_KEY
• S3_BUCKET your AWS S3 bucket path required
• S3_PREFIX path prefix in your bucket
• S3_REGION the AWS S3 bucket region (default: us-west-1)
• S3_ENDPOINT the AWS Endpoint URL, for S3 Compliant APIs such as minio⁠ (default: none)
• S3_S3V4 set to yes to enable AWS Signature Version 4, required for minio⁠ servers (default: no)
• SCHEDULE backup schedule time, see explanations below

⁠Automatic Periodic Backups

You can additionally set the SCHEDULE environment variable like -e SCHEDULE="@daily" to run the backup automatically.

More information about the scheduling can be found here⁠.

⁠Restoring a backup

Run restore.sh inside the container, passing either the name of a database (to restore its most recent backup) or an explicit S3 key:

$ docker exec -it <container> sh restore.sh dbname
$ docker exec -it <container> sh restore.sh dbname/2026/06/dbname.2026-06-21T00:52:18Z.sql.gz

The argument is treated as an S3 key if it ends in .sql, .sql.gz, .sql.enc, or .sql.gz.enc; otherwise it is treated as a database name and the most recent backup for that database is located automatically, using the same naming convention backup.sh writes (<db>/<year>/<month>/<db>[.version].<timestamp>.sql.gz[.enc]).

The backup is downloaded, decrypted (if ENCRYPTION_PASSWORD is set and the file ends in .enc), decompressed (if .gz), and restored into the database named by the leading path segment of the backup key with psql.