PDU Docs
PDU v3.0 Documentation
PostgreSQL Data Unloader ——
PostgreSQL Disaster Recovery ToolIntroduction
PDU is a disaster recovery tool designed for PostgreSQL versions 10-18. Created by Zhang Chen (Jay), this tool integrates data file mining and recovery of accidentally deleted, updated, or dropped data.
Design Philosophy
Why PDU was created and its core mission
Instant Recovery Guide
Quick recovery operations for various scenarios
Usage Guide
Configuration, parameters, and detailed usage
FAQ
Frequently asked questions and answers
PDU assembles all the core capabilities for PostgreSQL data recovery, providing comprehensive solutions from extracting data from corrupted files, recovering accidentally deleted/updated records, to restoring dropped tables without backups.
Core Capabilities
Deleted/Updated Data Recovery
Restores original data from DELETE/UPDATE operations using archived WALs
Direct File Extraction
Extracts data from files when databases won't start
Dropped Table Restoration
The only publicly documented practical method to restore dropped tables without backups
Safe & Read-only
All operations are read-only, never modifying original data files
Supported Data Types
Numeric & Serial
Text & Binary
Date & Time
JSONB & XML
Arrays
UUID & Boolean
Geography & Geometry
Unsupported types: enumeration types, text search vectors, composite types, and range types