Implementing Continuous Deployment with Jenkins & Gerritt
summary
Conference experience moves from slow brittle releases to a gated CI pipeline using Git, Gerrit and Jenkins that runs pretested commits via the Gerrit trigger plugin, enforces +2 approvals, and reports back into Gerrit. Tests run on many Jenkins slaves provisioned from OpenStack and configured with masterless Puppet; Capistrano handles deploys and selenium smoke tests validate test environments. Results: daily deploys, 2% failed deployments and typical failure stacks of ~14 commits, with developers iterating by squash with git rebase to keep changes small.
AI-written summary · about
chapters (22)
- 0:00Start
- 1:10Hacking with Ruby
- 1:37Let's talk about: continuous deployment
- 3:37Continuous Deployment is about
- 4:52Continuous Deployment is GOOD
- 5:56manual code review
- 6:17very little automation
- 8:47Step One: Automate Development
- 10:28Never stop automating.
- 11:03Step Two Better tools, processes
- 11:43centralized workflow
- 12:02lieutenants workflow
- 13:52Gerrit Trigger plugin
- 20:42Recap
- 21:25Step Three Automate Everything Else
- 21:42Deploying test environments
- 22:12New kinds of tests
- 24:14Step Four More Robots!
- 25:01OpenStack and the jclouds plugin
- 26:16nobody's perfect
- 26:52No production acceptance testing
- 27:23Tests are slow
related talks
01h 12mPitfalls of Continuous Deployment
46mContinuous Delivery On Premise: Gerrit, Jenkins and SonarQube ...
29mFrom scheduled builds to commit triggered pipelines
27mAutomation, Innovation and Continuous Delivery
42mZero to Jenkins: Automatic builds + deploys
36mUsing Jenkins to Build, Deploy, and Automate Projects
01h 12mPitfalls of Continuous Deployment
04mContinuous Delivery with Jenkins' New Declarative Pipeline & Artifactory