Policy-Based Volume Snapshots Management in Kubernetes
summary
Implements a Kubernetes VolumeSnapshot CRD plus out‑of‑tree controllers to create/delete snapshots and provision volumes from them (demo shows automated snapshots every 10s then 5s with 20–50s retention), currently working with GCE PD, RBD and HostPath drivers and moving toward CSI. Adds a SnapshotPolicy via MetaController CRD that periodically creates snapshots and prunes old ones; proposes an in‑place restore API that swaps a PVC to a newly provisioned PV to minimize pod downtime.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:10Overview
- 1:59Why Snapshot
- 4:37Kubernetes Volume Concepts
- 7:37Volume Snapshot Design
- 9:26Volume Snapshot Creation
- 10:22API Design
- 11:24Snapshot File
- 12:20Create New Volume
- 13:44Auto Snapshot API
- 14:25Snapshot Policy Controller
- 16:59Snapshot Policy Type
- 17:48Policy Controller
- 19:27Demo
- 22:53Restore Workflow
- 27:09Summary
- 28:14Questions
related talks
58mProtecting Stateful Workloads With CSI Snapshot
33mConsistent Volume Group Snapshots, Unraveling the Magic
35mBeyond Storage Management
25mIntro & Deep Dive: Kubernetes Data Protection WG
30mKubernetes Data Protection WG Deep Dive
37mKubernetes Storage 101
28mKubernetes Data Protection WG Deep Dive
31mKubernetes Data Protection WG Intro & Deep Dive