Inside Vitess: Building Resilient, Distributed Databases
summary
Vitess is presented as a Kubernetes-ready, MySQL wire‑compatible sharding platform with automatic failover (VReplication/Vtctld), online schema changes, two‑phase commit for cross‑shard transactions, and snapshot or object‑store backups (S3 plugin supported). Demo shows four geo shards: Americas, Europe, APAC, MEA with shard key mapping sending queries to one region only, and recommends 250 GB => 15 min recovery per shard to bound restore time. Query path is VtGate → vttablet → MySQL, and operations include resharding with near‑zero downtime and pluggable backup/restore.
AI-written summary · about
chapters (10)
- 0:00Vitess introduction
- 2:04Vitess vs single-node MySQL
- 5:54Vitess operations beyond sharding
- 7:56Advanced Vitess features
- 10:42Geosharding for global apps
- 12:54Vitess architecture deep dive
- 17:08Demo: geosharding setup
- 22:12Query routing & data residency
- 27:46Resources & community
- 28:31Q&A: backups & active-active
