From b2ba4181cd31122ee793a02f5381035beeeb9fbd Mon Sep 17 00:00:00 2001 From: Michael Thomson Date: Thu, 30 Jul 2026 14:38:27 -0400 Subject: [PATCH] nvidia plugin to operator --- .../nvidia-device-plugin/release.yaml | 27 -------------- .../nvidia-device-plugin/repository.yaml | 8 ---- .../nvidia-device-plugin/runtime-class.yaml | 5 --- .../nvidia-gpu-operator/release.yaml | 37 +++++++++++++++++++ .../nvidia-gpu-operator/repository.yaml | 8 ++++ .../namespace-nvidia-device-plugin.yaml | 8 ---- .../namespace-nvidia-gpu-operator.yaml | 6 +++ 7 files changed, 51 insertions(+), 48 deletions(-) delete mode 100644 infrastructure/controllers/nvidia-device-plugin/release.yaml delete mode 100644 infrastructure/controllers/nvidia-device-plugin/repository.yaml delete mode 100644 infrastructure/controllers/nvidia-device-plugin/runtime-class.yaml create mode 100644 infrastructure/controllers/nvidia-gpu-operator/release.yaml create mode 100644 infrastructure/controllers/nvidia-gpu-operator/repository.yaml delete mode 100644 infrastructure/namespaces/namespace-nvidia-device-plugin.yaml create mode 100644 infrastructure/namespaces/namespace-nvidia-gpu-operator.yaml diff --git a/infrastructure/controllers/nvidia-device-plugin/release.yaml b/infrastructure/controllers/nvidia-device-plugin/release.yaml deleted file mode 100644 index 4c8dc2e..0000000 --- a/infrastructure/controllers/nvidia-device-plugin/release.yaml +++ /dev/null @@ -1,27 +0,0 @@ -apiVersion: helm.toolkit.fluxcd.io/v2 -kind: HelmRelease -metadata: - name: nvidia-device-plugin - namespace: nvidia-device-plugin -spec: - chart: - spec: - chart: nvidia-device-plugin - version: 0.17.x - sourceRef: - kind: HelmRepository - name: nvdp - interval: 15m - releaseName: nvidia-device-plugin - values: - runtimeClassName: nvidia - config: - default: time-slicing - map: - time-slicing: | - version: v1 - sharing: - timeSlicing: - resources: - - name: nvidia.com/gpu - replicas: 5 diff --git a/infrastructure/controllers/nvidia-device-plugin/repository.yaml b/infrastructure/controllers/nvidia-device-plugin/repository.yaml deleted file mode 100644 index 4735cb9..0000000 --- a/infrastructure/controllers/nvidia-device-plugin/repository.yaml +++ /dev/null @@ -1,8 +0,0 @@ -apiVersion: source.toolkit.fluxcd.io/v1 -kind: HelmRepository -metadata: - name: nvdp - namespace: nvidia-device-plugin -spec: - interval: 15m - url: https://nvidia.github.io/k8s-device-plugin diff --git a/infrastructure/controllers/nvidia-device-plugin/runtime-class.yaml b/infrastructure/controllers/nvidia-device-plugin/runtime-class.yaml deleted file mode 100644 index c26bd6d..0000000 --- a/infrastructure/controllers/nvidia-device-plugin/runtime-class.yaml +++ /dev/null @@ -1,5 +0,0 @@ -apiVersion: node.k8s.io/v1 -kind: RuntimeClass -metadata: - name: nvidia -handler: nvidia diff --git a/infrastructure/controllers/nvidia-gpu-operator/release.yaml b/infrastructure/controllers/nvidia-gpu-operator/release.yaml new file mode 100644 index 0000000..e96f07c --- /dev/null +++ b/infrastructure/controllers/nvidia-gpu-operator/release.yaml @@ -0,0 +1,37 @@ +apiVersion: helm.toolkit.fluxcd.io/v2 +kind: HelmRelease +metadata: + name: nvidia-gpu-operator + namespace: nvidia-gpu-operator +spec: + chart: + spec: + chart: gpu-operator + version: 26.x + sourceRef: + kind: HelmRepository + name: nvidia + interval: 15m + releaseName: nvidia-gpu-operator + values: + driver: + enabled: false + toolkit: + enabled: false + hostPaths: + driverInstallDir: "/usr/local" + devicePlugin: + config: + create: true + name: "time-slicing-config" + default: "timeslicing" + data: + timeslicing: |- + version: v1 + flags: + migStrategy: none + sharing: + timeSlicing: + resources: + - name: nvidia.com/gpu + replicas: 5 diff --git a/infrastructure/controllers/nvidia-gpu-operator/repository.yaml b/infrastructure/controllers/nvidia-gpu-operator/repository.yaml new file mode 100644 index 0000000..0a0dbe9 --- /dev/null +++ b/infrastructure/controllers/nvidia-gpu-operator/repository.yaml @@ -0,0 +1,8 @@ +apiVersion: source.toolkit.fluxcd.io/v1 +kind: HelmRepository +metadata: + name: nvidia + namespace: nvidia-gpu-operator +spec: + interval: 15m + url: https://helm.ngc.nvidia.com/nvidia diff --git a/infrastructure/namespaces/namespace-nvidia-device-plugin.yaml b/infrastructure/namespaces/namespace-nvidia-device-plugin.yaml deleted file mode 100644 index 23d0bad..0000000 --- a/infrastructure/namespaces/namespace-nvidia-device-plugin.yaml +++ /dev/null @@ -1,8 +0,0 @@ -apiVersion: v1 -kind: Namespace -metadata: - name: nvidia-device-plugin - labels: - pod-security.kubernetes.io/enforce: privileged - pod-security.kubernetes.io/audit: privileged - pod-security.kubernetes.io/warn: privileged diff --git a/infrastructure/namespaces/namespace-nvidia-gpu-operator.yaml b/infrastructure/namespaces/namespace-nvidia-gpu-operator.yaml new file mode 100644 index 0000000..657ecee --- /dev/null +++ b/infrastructure/namespaces/namespace-nvidia-gpu-operator.yaml @@ -0,0 +1,6 @@ +apiVersion: v1 +kind: Namespace +metadata: + name: nvidia-gpu-operator + labels: + pod-security.kubernetes.io/enforce: privileged