Project-HAMi / HAMi

Heterogeneous AI Computing Virtualization Middleware
http://project-hami.io/
Apache License 2.0
956 stars 197 forks source link

Pod configuration scheduling failed #591

Closed armywork closed 2 weeks ago

armywork commented 2 weeks ago

Please provide an in-depth description of the question you have:

What do you think about this question?:

Environment: Ubuntu 20.04.2 LTS

调度

kubectl describe pod : no available node, all node scores do not meet

vgpu-scheduler-extender logs: request devices nums cannot exceed the total number of devices on the node." pod="default/test-6d5bc9d5dc-knrqr" request devices nums=16 node device nums=8 I1104 06:32:23.952296 1 score.go:225] "calcScore:node not fit pod" pod="default/test-6d5bc9d5dc-knrqr" node="k8s-master1"

pod.yaml resources: limits: cpu: "1" memory: 1Gi nvidia.com/gpu: 16 nvidia.com/gpumem: 20k requests: cpu: "1" memory: 1Gi nvidia.com/gpu: 16 nvidia.com/gpumem: 20k