feat: add nvidia time-slicing

[KCSesh]

Issue number:

Related:

• feat: add nvidia time-slicing bottlerocket-settings-sdk#62
• Allow changes to the NVIDIA device plugin configurations bottlerocket#2347

Description of changes:
Adding time-slicing configs to the core-kit and pointing to the latest settings-sdk.

Testing done:

1. Instance joined the cluster

NAME                                           STATUS   ROLES    AGE   VERSION
ip-XXXX.us-west-2.compute.internal   Ready    <none>   15h   v1.29.5-eks-1109419

2. Model Default:

bash-5.1# apiclient get settings.kubelet-device-plugin
{
  "settings": {
    "kubelet-device-plugins": {
      "nvidia": {
        "device-id-strategy": "index",
        "device-list-strategy": "volume-mounts",
        "device-sharing-strategy": "none",
        "pass-device-specs": true,
      }
    }
  }
}

3. Model Updates:

bash-5.1#: apiclient set settings.kubelet-device-plugins.nvidia.device-sharing-strategy=time-slicing settings.kubelet-device-plugins.nvidia.time-slicing.replicas=2 settings.kubelet-device-plugins.nvidia.time-slicing.rename-by-default=true settings.kubelet-device-plugins.nvidia.time-slicing.fail-requests-greater-than-one=true
bash-5.1# apiclient get settings.kubelet-device-plugin
{
  "settings": {
    "kubelet-device-plugins": {
      "nvidia": {
        "device-id-strategy": "index",
        "device-list-strategy": "volume-mounts",
        "device-sharing-strategy": "none",
        "pass-device-specs": true,
        "time-slicing": {
          "fail-requests-greater-than-one": true,
          "rename-by-default": true,
          "replicas": 2
        }
      }
    }
  }
}

4. Bounded check:

bash-5.1# apiclient set settings.kubelet-device-plugins.nvidia.time-slicing.replicas=1
Failed to change settings: Failed PATCH request to '/settings/keypair?tx=apiclient-set-Ne2SavnsTF7Fweq0': Status 400 when PATCHing /settings/keypair?tx=apiclient-set-Ne2SavnsTF7Fweq0: Unable to match your input to the data model.  We may not have enough type information.  Please try the --json input form.  Cause: Error during deserialization: integer out of range, expected it to be between 2 and 2147483647 at line 1 column 107
bash-5.1#

5. Files generated:

bash-5.1# cat /etc/nvidia-k8s-device-plugin/settings.yaml
version: v1
flags:
  migStrategy: "none"
  failOnInitError: true
  plugin:
    passDeviceSpecs: true
    deviceListStrategy: volume-mounts
    deviceIDStrategy: index
sharing:
  timeSlicing:
    renameByDefault: true
    failRequestsGreaterThanOne: true
    resources:
    - name: "nvidia.com/gpu"
      replicas: 2

nvidia-device-plugin[360007]:   "resources": {
nvidia-device-plugin[360007]:     "gpus": [
nvidia-device-plugin[360007]:       {
nvidia-device-plugin[360007]:         "pattern": "*",
nvidia-device-plugin[360007]:         "name": "nvidia.com/gpu"
nvidia-device-plugin[360007]:       }
nvidia-device-plugin[360007]:     ]
nvidia-device-plugin[360007]:   },
nvidia-device-plugin[360007]:   "sharing": {
nvidia-device-plugin[360007]:     "timeSlicing": {
nvidia-device-plugin[360007]:       "renameByDefault": true,
nvidia-device-plugin[360007]:       "failRequestsGreaterThanOne": true,
nvidia-device-plugin[360007]:       "resources": [
nvidia-device-plugin[360007]:         {
nvidia-device-plugin[360007]:           "name": "nvidia.com/gpu",
nvidia-device-plugin[360007]:           "rename": "nvidia.com/gpu.shared",
nvidia-device-plugin[360007]:           "devices": "all",
nvidia-device-plugin[360007]:           "replicas": 2
nvidia-device-plugin[360007]:         }
nvidia-device-plugin[360007]:       ]
nvidia-device-plugin[360007]:     }
nvidia-device-plugin[360007]:   }
nvidia-device-plugin[360007]: }

6. Time slicing on 1 instance with 1 GPU with 2 containers:

bash-5.1# apiclient set settings.kubelet-device-plugins.nvidia.device-sharing-strategy=time-slicing settings.kubelet-device-plugins.nvidia.time-slicing.replicas=2 
bash-5.1# apiclient get settings.kubelet-device-plugin
{
  "settings": {
    "kubelet-device-plugins": {
      "nvidia": {
        "device-id-strategy": "index",
        "device-list-strategy": "volume-mounts",
        "device-sharing-strategy": "time-slicing",
        "pass-device-specs": true,
        "time-slicing": {
          "replicas": 2
        }
      }
    }
  }
}

bash-5.1# cat /etc/nvidia-k8s-device-plugin/settings.yaml
version: v1
flags:
  migStrategy: "none"
  failOnInitError: true
  plugin:
    passDeviceSpecs: true
    deviceListStrategy: volume-mounts
    deviceIDStrategy: index
sharing:
  timeSlicing:
    renameByDefault: true
    failRequestsGreaterThanOne: true
    resources:
    - name: "nvidia.com/gpu"
      replicas: 2
bash-5.1#

➜  kubectl apply -f node-1-gpu-2.yaml
pod/gpu-test-2-pod-node-one created

 ➜  kubectl get pods
NAME                      READY   STATUS    RESTARTS   AGE
gpu-test-1-pod-node-one   1/1     Running   0          13m
gpu-test-2-pod-node-one   1/1     Running   0          3s

Terms of contribution:

By submitting this pull request, I agree that this contribution is dual-licensed under the terms of both the Apache License, version 2.0, and the MIT license.