Concept : Apply GPU acceleration to ML Node in k8s operator

Plugins Machine Learning
, , , ,
1

Versions (relevant - OpenSearch/Dashboard/Server OS/Browser):
OpenSearch 2.17.1
OpenSearch Operator 2.5

Describe the issue:

  1. Provision GPU-dedicated Worker Node(s) in the existing K8s cluster.
  2. Deploy OpenSearch ml-node(s) using nodeSelector in opensearch k8s operator.
  nodePools:
    - component: ml
      replicas: 3
      diskSize: "10Gi"
      nodeSelector: # <- assign ml node(ie. Pod) to GPU Worker Node 
      roles:
        - "ml"
      resources:
        requests:
        limits:
  1. Test whether ML Node does use GPU resource or not.

→ Is the above concept acceptable and valid ?

Actually, I can’t find any detailed reference about GPU acceleration for users who use k8s operator.