Backup & Recovery

Introduction

Backups save your business when something goes wrong — accidental DROP, ransomware, hardware failure. Untested backups are worse than no backups.

Understanding the topic

Core concepts to understand:

• Logical backups: pg_dump, mysqldump.
• Physical backups: file-level snapshot + WAL.
• PITR — point-in-time recovery using WAL.
• Off-site / encrypted / tested restores.
• Define RPO (data loss tolerance) and RTO (downtime tolerance).

Syntax reference

Visual workflow / architecture:

Daily logical backup ──▶  S3 (encrypted)
WAL streaming        ──▶  PITR storage
                          │
Restore drill (monthly):  rebuild + verify
                          │
                          ▼
                     RTO/RPO met

Real-world use

GitLab's 2017 incident proved that untested backups can mean total data loss. Every serious team runs monthly restore drills.

Best practices

• Automate backups + WAL.
• Test restores regularly — quarterly minimum.
• Define and meet RPO/RTO.

Hands-on exercise

Interview preparation — practice these questions:

• Q1. Logical vs physical backup.
• Q2. PITR — what is it?
• Q3. RPO vs RTO.
• Q4. Untested backup story.
• Q5. Restore drill frequency.