Build a Kubernetes Based Cloud Native Storage Solution From Scratch
summary
An open source distributed block storage system called Longhorn uses Kubernetes controllers, CRDs and the CSI plugin to run each volume as independent microservices (engine + replicas) written mostly in 30,000 lines of Go. It leverages modern SSDs, sparse files and Linux tooling, offers snapshots, S3 backups, cross‑cluster DR with user RTO/RPO, automatic replica rebuilds after node failure, and demonstrates live upgrade without impacting running volumes. Operational pain points include CSI/driver churn and informer cache races; plan moves engines to a daemonset to avoid pod limits and guarantee CPU.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:30Project Longhorn
- 1:10Why we need Longhorn
- 2:53Longhorn
- 4:50Longhorn Architecture
- 7:25Local Manager
- 9:22Controller Pattern
- 12:28Live Demo
- 20:38Upgrade Local Engine
- 21:20Demo
- 24:05Data Plan
- 24:47Resiliency
- 25:57Issues
- 29:19Upcoming Local Release
- 31:07Conclusion
- 31:45Audience Questions
related talks
48mIntro to Longhorn: Open Source Cloud-Native Distributed Block Storage Built On and For K8s
38mLonghorn: Intro, Deep Dive And Q+A
32mLonghorn: Intro, Deep Dive and Q&A
47mLonghorn: Intro, Deep Dive and Q+A
29mReliable and Cost-Effective: Open Storage Strategies for Kubernetes
32mUnderstanding Kubernetes Storage: Getting in Deep by Writing a CSI Driver
26mBest Practice for Cloud Native Storage in Finance Industry
46mUsing Kubernetes Local Storage for Scale-Out Storage Services