From 1890f434ce5e79729b54fa8295cee8ffe2b59046 Mon Sep 17 00:00:00 2001 From: Dan Manners Date: Sat, 11 Nov 2023 13:07:28 -0500 Subject: [PATCH] Adding GFD --- manifests/workloads/applicationset.yaml | 3 +++ manifests/workloads/nvidia/kustomization.yaml | 25 +++++++++++++++++++ 2 files changed, 28 insertions(+) diff --git a/manifests/workloads/applicationset.yaml b/manifests/workloads/applicationset.yaml index 0fd05c6c..3e994ce1 100644 --- a/manifests/workloads/applicationset.yaml +++ b/manifests/workloads/applicationset.yaml @@ -33,6 +33,9 @@ spec: - appName: node-feature-discovery namespace: kube-system targetRev: main + - appName: nvidia-device-plugin + namespace: kube-system + targetRev: main - appName: qemu-guest-agent namespace: kube-system targetRev: main diff --git a/manifests/workloads/nvidia/kustomization.yaml b/manifests/workloads/nvidia/kustomization.yaml index 56be4f0b..cc828957 100644 --- a/manifests/workloads/nvidia/kustomization.yaml +++ b/manifests/workloads/nvidia/kustomization.yaml @@ -10,3 +10,28 @@ helmCharts: version: 0.14.2 valuesInline: runtimeClassName: nvidia + gfd: + enabled: true + affinity: + nodeAffinity: + requiredDuringSchedulingIgnoredDuringExecution: + nodeSelectorTerms: + - matchExpressions: + # On discrete-GPU based systems NFD adds the following label where 10de is the NVIDIA PCI vendor ID + - key: feature.node.kubernetes.io/pci-10de.present + operator: In + values: + - "true" + - matchExpressions: + # On some Tegra-based systems NFD detects the CPU vendor ID as NVIDIA + - key: feature.node.kubernetes.io/cpu-model.vendor_id + operator: In + values: + - "NVIDIA" + - matchExpressions: + # We allow a GPU deployment to be forced by setting the following label to "true" + - key: "nvidia.com/gpu.present" + operator: In + values: + - "true" +