Developing Research Infrastructures the DevOps way
summary
A Virtual Research Environment for historians gets built and hardened: uses Jenkins and Puppet to produce Debian packages via fpm into an aptly repo, deployable to staging and production from Git branches; local development mirrors production with vagrant parity. The stack ingests about half a million datasets and runs ~1,000 users/month. Initial overhead and cultural friction give way to reproducible servers, coordinated weekly releases, and faster, safer emergency deploys.
AI-written summary · about
chapters (23)
- 0:00Introduction
- 1:10Sandre
- 1:52Research Infrastructures
- 2:33Humanities
- 3:17Centauri
- 5:28Project tree
- 9:13Machines
- 9:50Applications
- 10:45Reference URL
- 11:12Collaboration
- 12:27DevOps
- 16:01Application
- 16:36CCAN
- 17:59Communication
- 20:05Demo
- 24:12What we learned
- 30:53Fears
- 32:18Puppet code
- 34:21Sustainability
- 38:17Named Entity Recognition
- 40:39Changing the mindset
- 44:45Automated tests
- 46:43Acceptance by humanities researchers
related talks
33mDeveloping for Multi-Component Environments While Keeping Your Sanity
27mDevSecOps - Why Debian repositories are Mission-Critical!
36mDevOps for the Rubyist Soul
37mDevOps for the Rubyist Soul
19mDevOps Challenges: The Key Is Technical Excellence
33mDevOps and the Rebirth of Childhood Innocence- 33mDeveloping for Multi-Component Environments While Keeping Your Sanity
24mTaming Your Inner Cowboy Coder - A Simple And Sane DevOps Workflow