Skip to content

database-restore-drill.md

Latest commit

 

History

History
42 lines (28 loc) · 1.22 KB

database-restore-drill.md

File metadata and controls

42 lines (28 loc) · 1.22 KB
Raw

Runbook: Database Restore Drill

Scope

Use this to prove service-owned PostgreSQL backup and restore behavior for local portfolio evidence. Production restore should use managed RDS snapshots plus the same service ownership rules.

Drill Steps

  1. Start the local stack.
  2. Create a backup directory outside tracked source or under ignored backups/.
  3. Back up all service databases:
.\scripts\backup-postgres.ps1 -OutputDir .\backups -Databases all
  1. Restore one non-critical local database with explicit confirmation:
.\scripts\restore-postgres.ps1 -InputDir .\backups -Database devhire_job -ConfirmRestore
  1. Run post-restore verification:
.\scripts\dr-verify.ps1 -GatewayUrl http://localhost:8080
.\scripts\api-smoke.ps1 -GatewayUrl http://localhost:8080

Guardrails

  • Restore is not destructive unless -ConfirmRestore is provided.
  • Do not restore one service database from another service backup.
  • Do not commit backup files.
  • Record RPO/RTO and drill date in release evidence.

Expected Evidence

  • Backup files exist in ignored storage.
  • Restored service responds through Gateway.
  • Search and application flows still use service APIs, not cross-database reads.