Ceph CSI Driver: Bridging Containers and Ceph
summary
Ceph CSI implements gRPC CSI servers in Go to let Kubernetes (and other orchestrators) provision RBD block volumes, CephFS subvolumes and NFS exports (RBD, CephFS, NFS). It uses go-ceph connection pools and rados omap for request checkpointing, supports layering/async flattening, dm-crypt/LUKS keying, a pod network namespace file to keep mounts alive across container restarts, an RBD NBD healer, and CSI add-ons with CRDs for reclaim-space, network-fence and volume-replication. The system is production-ready and shows measurable latency and connection-reuse improvements.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 0:50Agenda
- 1:37Containers
- 2:52Storage Drivers
- 4:03Ceph CSI
- 7:57Standard Deployment
- 10:00RBD Driver
- 13:56Storage Class
- 15:52RBD
- 18:10Omap Regeneration
- 21:12CFS Driver
- 25:17NFS Driver
- 26:13CSI Addons
- 27:37CSI Operations
- 30:19Volume Replication
- 31:53Future Roadmap
- 34:27QA
related talks
27mCeph-CSI: The Engine Behind Persistent Storage in Kubernetes
32mUnderstanding Kubernetes Storage: Getting in Deep by Writing a CSI Driver
48mCSI in Kubernetes
01h 24mTutorial: Introduction to Using the Container Storage Interface (CSI) Primitives
44mContainer Management at Scale
27mCeph, the future of Storage
35mDeploy Ceph in Kubernetes: Rook Deep Dive
39mContainer Storage Interface: Present and Future