pg_back

Project Overview

pg_back is a dump tool for PostgreSQL that dumps all or some databases with globals at once in the format you want. Behind the scene, it uses pg_dumpall to dump roles and tablespaces, and pg_dump to dump each selected database to a separate file in the custom format.

Key Features

• Dump all or a list of databases, or exclude specific databases
• Include database templates
• Choose the dump format per database
• Limit dumped schemas and tables
• Dump databases concurrently
• Compute SHA checksum of each dump
• Pre-backup and post-backup hooks
• Purge based on age and number of dumps to keep
• Dump from a hot standby by pausing replication replay
• Encrypt and decrypt dumps
• Upload and download to S3, GCS, Azure, B2, or SFTP

Target Audience

• PostgreSQL database administrators
• DevOps engineers
• Teams who need automated backups

Quick Start

1. Download a compiled binary from Releases
2. Or install via Go: go install github.com/orgrim/pg_back@latest
3. Connect to PostgreSQL using -h, -p, -U, and -d
4. Specify output directory: -b /var/backups/postgresql

Core Advantages

• Parallel dumps for speed
• Multiple storage backends (S3, GCS, Azure, B2, SFTP)
• Encryption and checksums for data integrity
• Open-source and free (MIT)

System Requirements

• OS: Cross-platform
• Architectures: x86_64, arm64
• Additional Dependencies: pg_dump and pg_dumpall (PostgreSQL client tools)