README.md

s3pgbackups

PostgreSQL cluster backups to Amazon S3.

Overview

This application will backup your PostgreSQL cluster to Amazon S3, which includes:

• Separate SQL pg_dump of each database
• Compressed data (gzip)
• Automatically rotated backups
• Ability to exclude specific tables, or entire databases

Installation

Debian

$ curl -s https://packagecloud.io/install/repositories/gregarmer/packages/script.deb.sh | sudo bash

CentOS (or RHEL / other rpm based distro's)

$ curl -s https://packagecloud.io/install/repositories/gregarmer/packages/script.rpm.sh | sudo bash

Getting Started

1. Follow the installation steps above.

2. Configure ~/.s3pgbackups - something like this:

   {
     "aws_access_key": "UBGKJEBGKE56783JHVFW",
     "aws_secret_key": "webgrwebgjwbewegfkeg",
     "s3_bucket": "your.db.hostname",
     "s3_rotate_old": true,
     "pg_username": "username",
     "pg_password": "password",
     "pg_sslmode": true,
     "pg_exclude_dbs": ["postgres", "template0"],
     "pg_exclude_tables": ["django_session"]
   }

   Note: Running s3pgbackups without a config will create a blank config.

3. Run s3pgbackups

Usage

$ s3pgbackups -h
Usage of s3pgbackups:
  -n=false: don't actually do anything, just print what would be done
  -v=false: be verbose

You can use the -n parameter to put s3pgbackups into no-op mode, where no actions will actually be performed.

The -v parameter will make s3pgbackups be verbose about what is actually happening.

TODO

• Support for other RDBMS's
• Support for other targets, besides Amazon S3
• If we do the above, we should probably rename the project ?
• Using S3 behind a proxy - depends on this PR