You Don't Need a Database Backup Policy
summary
Claim: a backup policy alone is useless unless it’s driven by recovery requirements. Define RPO, RTO and retention, then pick methods that meet them: physical online backups plus continuous WAL archiving for PITR, logical dumps for object-level restores, or storage snapshots with coordinated volumes. Tools discussed include pgBackRest, Barman, pg_basebackup and pg_dump. Example timelines: HA failover in seconds, standalone restore ~45 minutes, table-repair via delayed replica using recovery_min_apply_delay can avoid full rollback. Test recoveries at least every six months.
AI-written summary · about
chapters (15)
- 0:00Intro
- 1:04What is a Database Backup Policy
- 2:03Why take backups
- 3:45What are your recovery requirements?
- 6:21Backup methods
- 12:38Backup tools
- 14:15Creating a disaster recovery policy
- 14:41Recovery requirements
- 15:38Responsibilities
- 15:59Recovery procedures
- 22:10Test each scenario
- 22:43Backup strategy
- 23:35Monitoring policy
- 24:35Ongoing testing & monitoring of your disaster recovery policy
- 26:59Conclusions
related talks
22mPostgreSQL Disaster Recovery Fundamentals
56mPostgreSQL Backup Strategies
50mPostgreSQL Backup and Recovery Methods
56mI Dropped My Database! Now What? - A Dive Into PostgreSQL Backup Using pgBackRest
59mPostgreSQL Don'ts and Dos
48mCTRL+Z for your Database : Mastering Backup and Recovery in PostgreSQL
49mTools for Creating PostgreSQL Backups
46mLock, Stock and Backup: Data Guaranteed