Sturdy Clouds Creating Resilient Software Defined Data Centers
summary
Argues that resilience requires concrete choices: deploy OpenStack and Kubernetes with repeatable automation, treat automation is everything, avoid exotic hardware and absurdly compact Ceph setups (three-node 80‑disk example warns of endless resyncs), run a small test cloud for feature rollouts, keep DR geographically separate from primary and follow disaster recovery != backup (backups must allow rollbacks, not only endless incrementals), separate OS/app/user data, target measurable SLAs (example: 99.9% API) and centralize logs/metrics.
AI-written summary · about
chapters (14)
- 0:00Start
- 5:42Why are we building redundancy?
- 7:49101 Indrastructure Best Practices
- 10:31Low Risk Deployments
- 12:51Geo Redundancy and DR
- 16:31Disaster Recovery Plan
- 18:42Protect your operation - Backup
- 20:04Open Source Recap
- 21:28Open Source Risks and Challenges
- 23:06Open Source Checklist plan
- 24:14Security
- 25:39Cloud Native Disaster Recovery
- 28:09Observability is key
- 29:25MIRANTIS
related talks
29mThe Future of the Cloud
22mResilience Is Not Enough
29mCeph, Xen, and CloudStack: Semper Melior
15mIf You Can, It Doesn't Mean You Should: Lessons From Terraforming Clouds
42mThe State of Open Source Cloud
41mA vision for the future
36mDesigning for failure - architecting fault-tolerant system
16mSAG Expert Interview #12: The Art of Resilient Software Architecture