The Topology Manager is a kubelet component that aims to coordinate the set of components that are responsible for these optimizations. After Topology Manager was introduced the problem of launching pod in the cluster where worker nodes have different NUMA topology and different amount of resources in that topology became actual. The Pod could be scheduled in the node where the total amount of resources is enough, but resource distribution could not satisfy the appropriate Topology policy.
When allocating CPUs for LSE/LSR Pods to implement static binding, NUMA Topology also needs to be considered.
More details please refer to #96
TODO:
[x] koordlet reports complete NUMA Node information
[x] koordlet supports NUMA Share Pool
[x] koord-scheduler supports NUMA Topology Alignment scheduling with CPU/Memory
What is your proposal:
The Topology Manager is a kubelet component that aims to coordinate the set of components that are responsible for these optimizations. After Topology Manager was introduced the problem of launching pod in the cluster where worker nodes have different NUMA topology and different amount of resources in that topology became actual. The Pod could be scheduled in the node where the total amount of resources is enough, but resource distribution could not satisfy the appropriate Topology policy.
When allocating CPUs for LSE/LSR Pods to implement static binding, NUMA Topology also needs to be considered.
More details please refer to #96
TODO: