Deep Dive: Kubernetes/Kubespray (Cluster Lifecycle) SIG
summary
Kubespray uses Ansible to deploy production Kubernetes on bare metal and clouds, focusing on upgrades, HA and plugin choice; deployments vary from a 5‑minute single‑node demo to common 30‑node runs and at 500 nodes a two‑run strategy cuts hours (three vs five). Features include tags to rerun only failed phases, a download‑only role to populate a local registry, prebuilt AMI/VM images and experimental Mitogen integration claiming ~3x faster runs; etcd snapshots and kubeadm control‑plane join improvements are already used.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 0:35Project Principle
- 4:10Continuous Integration
- 4:31Contribution to Keepspray
- 5:19Speed
- 6:09Show Hands
- 6:44Deployment
- 7:41Tags
- 13:27Reducing Tasks
- 15:00Wrapping
- 16:42Cubespray
- 21:19Disaster recovery
- 26:48Immutable deployments
- 28:43Selfjoin
- 30:53Operators
related talks
32mDeep Dive: Kubespray (a sig-cluster-lifecycle's project)
32mDeep Dive: Kubespray (a sig-cluster-lifecycle's project)
36mKubespray Intro
44mSIG Cluster Lifecycle: What's New in Kubespray
12mKubespray: Driving Cost-Efficiency for AI on Kubernetes
40mBuilding Custom GPU Clusters at Scale: Using Kubespray To Create High-Performance AI Infrastructure
30mDeploying Kubernetes Without Scaring Off Your Security Team
43mDeep Dive: SIG Cluster Lifecycle