Backup and Restore: Discover Best Practices for a Distributed Database Backup Strategy

on YouTube ↗27m

" summary

YugabyteDB runs online backups that are topologically independent, copying tablet-level snapshot hard links in parallel to S3/GCS/Azure or NFS with SHA-256 checksums; YCQL supports table/keyspace backups while YSQL is database-level. Restores map tablet IDs (example: RF3 → RF1) and replay data; PITR uses in-cluster snapshots plus PITR with MVCC (snapshots every 24h, three-day retention yields continuous history and microsecond RPOs). Backups 2.0 adds a per-node sidecar and delta uploads, claiming 4-5x faster backups and lower network/storage costs.

AI-written summary · about

" chapters (9)

1. 0:00Start
2. 0:32What Is the Purpose of Backup
3. 6:50Backing Up a Database
4. 8:06Create a Backup
5. 13:44Restoring a Database
6. 17:42Point in Time Recovery
7. 21:27Point in Time Recovery Management
8. 22:09Backup Studio
9. 25:51Distributed Snapshots

" related talks

• 57mNo Downtime, No LimitsKarthik Rangathan · Distributed SQL Summit Asia 2024
• 27mIntroduction to YugabyteDB – Design and ArchitectureKarthik Ranganathan · Distributed SQL Summit Asia 2021
• 26mIntroducing YugabyteDB for PostgreSQL UsersBo Peng · Distributed SQL Summit Asia 2023
• 32mDay 2 Operations Best PracticesTaylor Mull, John Forrest · Distributed SQL Summit Asia 2021
• 25mYugabyteDB Technical OverviewTOMOHIRO ICHIMURA · Distributed SQL Summit Asia 2022
• 27mIntroduction to YugabyteDB: Design and ArchitectureKarthik Ranganathan · Distributed SQL Summit 2020
• 32mArchitectural Flexibility and Geo-Specificity with Cluster-to-Cluster ReplicationHariprasad Mohanarangan · Distributed SQL Summit Asia 2025
• 30mBuilding Cloud Native Applications on Distributed SQL PlatformSHIGERU URUSHIBARA · Distributed SQL Summit Asia 2023