Testing Ceph: Status, Development, & Opportunities
summary
Ceph testing runs on the teuthology framework driving a 250+ server lab and thousands of machine‑hours of daily jobs; it combines combinatorial suite fragments, archived logs and OSD thrashers (chaos monkeys) and catches many regressions. Current tooling uses a primitive Beanstalk queue scheduler, lacks systemd and Kubernetes integration, has no automated performance trend analysis, and misses large‑scale and orchestrator tests — most pieces work but several gaps need code and infra changes.
AI-written summary · about
chapters (12)
- 0:00Introduction to Ceph testing
- 1:01History of teuthology framework
- 7:05Sepia lab and community testing
- 10:40QA suites and thrashers
- 17:08Current test coverage strengths
- 18:52Gaps in unit and integration tests
- 21:20Performance and scale testing gaps
- 23:47Deployment tool integration issues
- 24:31teuthology framework weaknesses
- 28:22Trend analysis and test tracking
- 30:10Suite explosion and scheduling problems
- 33:18Modernizing teuthology with Kubernetes
related talks
38mTesting Ceph: Pains & Pleasures
13mKeynote: Getting to Know Ceph
24mThe Ceph Community: A Driving Force for Project Evolution
29mCeph and the future of Open Source Storage
29mCeph's journey at SUSE
01h 02mIntroduction to Ceph and Architectural Overview
39mCeph Scientific Users Working Group BoF
04mKeynote: Building Together: Open Storage for Today and Tomorrow
