Two Years with Vitess: How JD.com Runs the World's Largest Vitess
summary
JD.com runs Vitess on Kubernetes at massive scale—over 20,000+ DB tablets and ~3,000 vttablets—using binlog streaming, etcd watches and a homegrown OFC orchestrator to automate master failover, resharding in seconds and topology updates. Operational work includes custom protocol patches, two-phase commit to avoid partial commits, a JTransfer binlog adapter for migration, and ML-driven resource scheduling to cut machine and dev cost. Many features required engineering patches; the system is production-proven but still evolving.
AI-written summary · about
chapters (10)
- 0:00Intro
- 3:30Dividing the database
- 6:30Adding data
- 9:30Charting system
- 16:00Internal Architecture
- 17:58Key Systems
- 21:06Business Launching
- 22:27Table Creation
- 25:49Advices
- 28:56Future Plans
related talks
35mTwo Years with Vitess: How JD.com Runs the World's Largest Vitess
36mRunning Vitess on Kubernetes at Massive Scale: JD.com Case Study
43mVitess and a Kubernetes Operator
44mIntro: Vitess
40mVitess: Intro and New Features
45mDeep Dive: Vitess
31mVitess: Running MySQL with confidence on Kubernetes
40mVitess: Introduction And New Features