Take a BlaBlaCar to the stars with rkt
summary
Company moves 300 bare‑metal servers to CoreOS and rkt for production, building about 400 images and running ~4,000 containers with ~80% of production migrated by July 2015. Orchestration uses fleet plus a homegrown layer called GGN to deploy pods; images are built with digger and pods use service discovery via go‑synapse/go‑nerve. Databases (Cassandra, Galera MySQL, Redis, Elasticsearch) run in containers with app‑level replication; outcome is stable production deployment that tolerates random hardware reboots.
AI-written summary · about
chapters (26)
- 0:00Introduction
- 1:23Infrastructure
- 2:15Tech Revolution
- 4:51Network
- 5:55Rebooting
- 7:53Serviceoriented
- 8:50Containers
- 11:54Rocket
- 18:14buildrun
- 19:13builder
- 21:03templating
- 22:09run levels
- 22:43Testing
- 23:26Summary
- 24:09Fleet Feed
- 25:10Directory Structure
- 26:27Override Attributes
- 27:09Service Update
- 27:58Service Discovery
- 28:43Airbnb Smart Stack
- 30:35barometer servers
- 31:48Orchestration
- 33:37Questions
- 36:51Future of rkt
- 38:16Storage
- 42:35Uniformization
related talks
39mSuper Fast Real-time Data Processing on Cloud-Native Architecture
08mContainerization – It’s Not Just for CPUs Anymore: Introducin...
18mKeynote: Mind the Gap: Bridging Cloud Native Innovation with Real-World Use
47mBuilding sustainable architectures with Kubernetes
29mThe Progress for Cloud Computing on Arm architecture
31mScaling a Distributed Container Management System: Diego Performance
37mCRI: The Second Boom of Container Runtimes - Harry Zhang & Xu Wang, HyperHQ
15mDemo: Resource(Cost) optimization on planet scale Kubernetes clusters.