SWITCH: Operations, Data Management and Automation
summary
SWITCH runs three Ceph generations and claims OSv3 replaces brittle scripts with a GitOps pipeline using Terraform, GitLab CI/CD and an OSv3 Ansible collection (built atop cephadm) to provision ~20 PB clusters with full IPv6 and multisite metadata. A single inventory plus a short YAML can bootstrap clusters in minutes and create multisite S3 replication automatically; migration tools move data off legacy Clouidian clusters with minimal user impact while pipelines and diff outputs ensure declared state matches running config.
AI-written summary · about
chapters (11)
- 0:00Introduction to SWITCH
- 1:03Switch storage solutions overview
- 3:11Legacy storage systems challenges
- 5:20GitOps approach for infrastructure
- 7:26Ansible collection for Ceph ADM
- 10:18Configuration management repository
- 12:04Role-based automation examples
- 14:50Multi-site cluster deployment
- 17:14Switch Cloud portal demo
- 20:24Pipeline and GitLab integration
- 25:56Live demo and Q&A
related talks
40mJuggling Petabytes: Managing Ceph at Scale with Ceph-ansible
21mBack to the original mission: an open cloud operating system
16mTips and Tricks for Managing and Administering Ceph Clusters
44mContainer Management at Scale
29mTowards Sustainable Multi-Cluster Management in Real-World Academic Infrastructure
36m200 Clusters vs 1 Admin
17mKeynote: Ceph Journey - A Perspective
20mEnterprise-Ready Ceph: Simplifying Storage Management, Operations, and Deployment