Scalable Deployments...
summary
Claim: replace Capistrano + SSH + rsync with Mamiya and push a built package to a master/agent gossip cluster that prepares, then switches symlinks; move expensive steps like bundle install and assets:precompile into a prior prepare phase. 140 servers deploy in about 30 seconds for 140 servers (reported ~16× speedup), support automatic fetch/prepare/release, HTTP API for agents, and one-command rollback that auto-prepares missing packages.
AI-written summary · about
chapters (11)
- 0:00Start
- 1:50RULES ON DEPLOYMENT
- 2:17Our deployment was
- 3:12How long is spent for deployment?
- 5:13How long time spent for deployment?
- 6:27IMPROVEMENT PLANS
- 6:52CREATE NEW TOOL!
- 7:49GOSSIP PROTOCOL?
- 9:11Terminologies & Concept
- 10:40Steps
- 20:41Future Plans
related talks
47mDeploying Like a Boss: From 2 Servers to 20,000
45mLarge-Scaled Deploy Over 100 Servers in 3 Minutes
36mContinuous Deployment with Rails
46m10+ Deploys Pe
40mFrom Push to Pull: Managing Mutable Infrastructure at a Global Scale
38mDeploying your apps... the right way!
46mRuby in the Cloud
24mContinuous Deployment at Scale
