Lightning Talk: Orchestrating Kubernetes Clusters on HPC Infrastructure
Lightningat Kubernetes Batch + HPC Day Europe 2023 · on Tuesday, 2023-04-18· at 16:30· Hall 7, Room A | Ground Floor | Europe Complex·
summary
CSCS builds centrally managed Kubernetes across bare metal, virtual and the Alps HPE Shasta HPC system using Rancher, spins on-demand clusters via Harvester into VLANs, and deploys everything with a ArgoCD app-of-apps GitOps flow. Diskless compute nodes force persistent state onto Ceph RBD/CephFS and S3 backups. Clusters run k3s or rke, Cilium replaces Calico, and the team reports storage I/O and scaling hotspots (about 100 Alps nodes supporting WCG) driving exploration of alternative storage backends and scheduler approaches.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 0:18About me
- 0:40Alps HPC system
- 1:32Kubernetes at CSC
- 3:05Challenges
- 4:06Argo CD
- 4:32Conclusion
- 5:45Discless nodes
related talks
10mLightning Talk: Creating A Scalable Foundation for Kubernetes Mult...
05mProject Lightning Talk: Multi-Cluster Configuration Management with KubeStellar
21m07 Lightning talks
06m⚡ Lightning Talk: Colocating Data Workloads and Web Services on Kubernetes to Improve Resource Utilization
29mTowards Sustainable Multi-Cluster Management in Real-World Academic Infrastructure
41mBuilding & Operating a Large-scale HPC AI Cluster on Kubernetes
19mKeynote: Building Bridges: Cloud Native and High Performance Computing
09m⚡ Lightning Talk: Advanced GPU-Orchestrated Workflows and HPC Integrations on K8s for Distributed AI/ML at Scale