Keynote: Debunking the Myth: Kubernetes Storage is Hard
summary
Claims that Kubernetes storage is hopelessly hard get narrowed into concrete layers: select storage vs deploy vs integrate vs consume. Block and file storage work via a CSI driver plus StorageClass and PersistentVolumeClaim, so swapping providers only needs a new driver and class. Complex stateful apps use Operators; Rook runs Ceph as an example. Managed services (Cloud SQL, GCE PD) are fine to consume from Kubernetes. Kubernetes automates provisioning, scheduling and moving volumes when nodes fail.
AI-written summary · about
chapters (22)
- 0:00Introduction
- 0:22Benefits of Kubernetes
- 0:44Stateful workloads
- 1:13Options
- 1:34Storage is hard
- 2:16What Kubernetes is responsible for
- 2:50Separating concerns design principle
- 3:30Storage problems
- 5:13Storage categories
- 5:50Requirements
- 6:33Make a decision
- 6:58Managed solutions
- 8:37Kubernetes Operators
- 10:24SoftwareDefined Storage Systems
- 12:08The Next Challenge
- 13:19Storage System
- 13:34Data Services
- 14:45Portability
- 15:25Availability
- 15:43Provisioning
- 18:06Storage is complicated
- 18:31Kubernetes abstractions
related talks
35mKubernetes Storage Lingo 101
49mDevOps: Kubernetes Storage - Myths, Facts and Tips
05mSponsored Keynote: Keep Kubernetes Caffeinated
05mWielding the Power & Potential of Kubernetes for Mere Mortals...
17mKeynote: Kubernetes: Finally...A True Cloud Platform
37mYou Have Stateful Apps - What if Kubernetes Would Also Run Your Storage?
32mUnderstanding Kubernetes Storage: Getting in Deep by Writing a CSI Driver
36mTips and tricks for Kubernetes storage