The Multicluster Toolbox
summary
Shows how an actual system is built and run across three GKE contexts (cluster0, cluster1, cluster2): a controller that watches Foo in one cluster and creates Deployments in another, using multicluster-controller plus a declarative bootstrap and multicluster-service-account to import tokens so pods get kubeconfigs. OwnerReferences are extended with cluster and namespace, garbage collection uses a poor-man's controller garbage collection via finalizers, and after one manual bootstrap the demos create and delete cross‑cluster resources successfully.
AI-written summary · about
chapters (10)
- 0:00Start
- 0:10The Multi Cluster Toolbox
- 1:30What Does Federation V2 Solve
- 5:24The Sample Controller Project
- 6:41Custom Resource
- 14:11Owner References
- 21:26Modes of Deletion in Kubernetes
- 21:59Garbage Collection Controller
- 29:40Service Account Import
- 32:07Custom Resource Definition
related talks
38mMulticluster Kubernetes, today and tomorrow
06mKeynote: From One to Many, the Road to Multicluster
35mDynamic Multi-Cluster Controllers With Controller-runtime
46mSIG Cluster Lifecycle (Cluster API)
34mMulticluster Ingress Powered by Kubernetes Cluster Registry - Greg Harmon & Nikhil Jindal
50mMulti-Cluster Kubernetes and Service Mesh Patterns
01h 24mTutorial: So You Want To Develop a Clust...
40mKubernetes is Your Cloud Control Plane
