CI and automated testing in Fedora: where we're at
summary
Fedora QA runs two complementary systems: OpenQA for high‑level functional checks that emulate a human on a VM and Fedora CI (Testing Farm + Jenkins) for contributor tests and gating. OpenQA runs about 200 compose tests each night, has run 3 million tests since 2015, and catches dozens of release‑blocking bugs; update testing runs 50–60 critical‑path tests per update. Results land in Bodhi, can gate updates, and include screenshots, logs and crash dumps for triage.
AI-written summary · about
chapters (23)
- 0:00Fedora CI and automated testing
- 1:06Four testing levels
- 2:12Manual testing limits
- 5:46Why automate Fedora
- 6:07Fedora CI vs openQA
- 9:36openQA automates QA
- 10:38openQA tests Fedora composes
- 13:13openQA install and upgrade tests
- 16:46Server tests and updates
- 18:29openQA for Fedora stability
- 19:53Bugs caught by openQA
- 21:16openQA open source service
- 22:21Fedora CI workflows
- 23:03Running Fedora CI tests
- 25:54Packet workflow for CI
- 26:34Other Fedora test systems
- 28:00Bodhi test results
- 30:07Gating updates and Rawhide
- 31:54Handling packager failures
- 32:57Test results integration
- 33:39Summary of Fedora CI
- 34:20Future of openQA and CI
- 35:50Q&A
related talks
48mFedora Automated Testing State of the Art: Not Quite CI for Distributions, but...
29mTowards stronger checks on dependencies and compose inputs in Fedora
21mThe emerging of the virtual QA team for Linux kernel
38mopenQA - end to end testing of operating systems
27mPlugging into the Red Hat kernel CI ecosystem
38mQuality Beyond Kernelci.Org in Upstream Linux for TI SoCs
34m7 Ways to Make Kernel Developers Like Cookies
24mContinuously Integrating the Upstream Linux Kernel on Hardware