Provisioning the Distributed Database TiDB on Kubernetes
summary
TiDB is a distributed SQL database with three core components (PD metadata store, TiKV storage using RocksDB+Raft, and a MySQL‑compatible SQL layer) and is provisioned on Kubernetes via a custom operator and Helm charts. The implementation adds support for local persistent volumes (local SSDs), multi‑tenant cluster management, automated failover and rolling upgrades, and a scheduler extender that enforces topology and virtual‑resource reservations to avoid pending pods. Deployed clusters with 3+ replicas run successfully in production and handle scaling and failover.
AI-written summary · about
chapters (10)
- 0:00Intro
- 1:09TIDB architecture
- 4:06TIDB ecosystem
- 7:49Cloud-Native applications
- 9:06Kubernetes: standard platform
- 11:11Kubernetes: powerful extensibility
- 13:13Features
- 16:21How we manage state
- 19:00Custom controller
- 25:39How we schedule stateful app
related talks
29mDistributed SQL How TiDB Modernized Data Management
47m10 Reasons Why the TiDB Database Should Power Your Next Development Project
29mDistributed SQL: How TiDB Modernized Data Management
40mTiKV Best Practices
46mTiDB 101
22mElastic Scheduling with TiKV
49mThe Paradigm Shift - Understanding the Move To Distributed Databases
40mTiKV Best Practices