Drupal HA: a story from the trenches
summary
Claims a reproducible high‑availability Drupal platform built with packaged as RPMs, pulp + nexus for internal repos and a Jenkins pipeline that promotes artifacts only if built from in‑house resources. Uses Puppet for infrastructure as code, HAProxy + Apache frontends, master‑master MySQL across DCs (fixing geospatial latency), Redis for caches and explicit cache‑flush orchestration after discovering racey Twig caches and DB overload from cache writes. The setup reduces emergency manual deploys and separates code deploy from config toggles, and it works for their teams.
AI-written summary · about
chapters (15)
- 0:00Intro
- 2:17Use Version Control for everything
- 3:13Deploy artifacts, not code
- 6:13Create artifacts from in-house resources
- 9:22The only way to get to prod are the (CI/CD) pipelines
- 10:37Config != Code
- 11:51Config management for all other things
- 12:57Measure all the Things
- 13:50N 1 for all services
- 16:00Shared Storage and Cache
- 17:03Geo-spatial Considerations
- 18:21Master - Master Replication woes
- 18:53Redis Cache
- 19:59Composer optimizations
- 20:54Post-install Optimizations
related talks
50mAutoscaling, Rolling Deploys with Drupal on AKS
01h 10mHosting Drupal on AWS
43mAutomated server management with open source tools
24mTaming Your Inner Cowboy Coder - A Simple And Sane DevOps Workflow
38mWhat's Really to Blame?
33mDeveloping for Multi-Component Environments While Keeping Your Sanity
01h 02mThe Road to continuous deployment
45mAgile Development of CakePHP Web Applications in a Hybrid Kubernetes Cluster