Deploy, Scale and Sleep at Night with JRuby
summary
JRuby on the JVM replaces per-process MRI pools with a single multi‑threaded instance, cutting memory use and GC overhead and keeping throughput rising where MRI servers hit a pool-size plateau (benchmarks show MRI flattening while JRuby keeps scaling). Deployment options: warbler war files for hot deploy and consistency, Trinidad for Capistrano‑style deploys, and TorqueBox for full app‑server features like high‑availability singletons and clustering. Run jr-lint to find JRuby compatibility issues; real users scale analytics by adding TorqueBox nodes.
AI-written summary · about
chapters (10)
- 0:00The architecture of scale
- 2:50Ruby on the JVM
- 4:45Challenges for sysadmins
- 6:58MRI architectural limitations
- 9:01Benefits of JRuby threading
- 13:02Preparing for deployment
- 15:41Warbler deployment strategy
- 19:45Trinidad for MRI users
- 21:11Power of TorqueBox
- 24:54Choosing your deployment path
