From 3cf940d273affb1024b62a5c196850d23bbac0ca Mon Sep 17 00:00:00 2001
From: aggarwalakshun <aggarwalakshun@gmail.com>
Date: Sat, 25 Oct 2025 14:55:10 +0530
Subject: [PATCH] Add Nvidia GPU operator Helm configuration files

---
 .../gpu-operator/gpu-operator-configmap.yml   | 18 +++++++++++
 .../gpu-operator/gpu-operator-release.yml     | 31 +++++++++++++++++++
 .../helm/gpu-operator/gpu-operator-repo.yml   |  9 ++++++
 3 files changed, 58 insertions(+)
 create mode 100644 clusters/default/helm/gpu-operator/gpu-operator-configmap.yml
 create mode 100644 clusters/default/helm/gpu-operator/gpu-operator-release.yml
 create mode 100644 clusters/default/helm/gpu-operator/gpu-operator-repo.yml

diff --git a/clusters/default/helm/gpu-operator/gpu-operator-configmap.yml b/clusters/default/helm/gpu-operator/gpu-operator-configmap.yml
new file mode 100644
index 0000000..7d89550
--- /dev/null
+++ b/clusters/default/helm/gpu-operator/gpu-operator-configmap.yml
@@ -0,0 +1,18 @@
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: time-slicing-config
+  namespace: gpu-operator
+data:
+  any: |-
+    version: v1
+    flags:
+      migStrategy: none
+    sharing:
+      timeSlicing:
+        resources:
+        - name: nvidia.com/gpu
+          replicas: 4
+
+# remember to patch the cluster policy to use this configmap
+# kubectl patch clusterpolicy/cluster-policy -n gpu-operator --type merge -p '{"spec": {"devicePlugin": {"config": {"name": "time-slicing-config", "default": "any"}}}}'
diff --git a/clusters/default/helm/gpu-operator/gpu-operator-release.yml b/clusters/default/helm/gpu-operator/gpu-operator-release.yml
new file mode 100644
index 0000000..a7d924b
--- /dev/null
+++ b/clusters/default/helm/gpu-operator/gpu-operator-release.yml
@@ -0,0 +1,31 @@
+---
+apiVersion: helm.toolkit.fluxcd.io/v2
+kind: HelmRelease
+metadata:
+  name: gpu-operator
+  namespace: gpu-operator
+spec:
+  interval: 24h
+  chart:
+    spec:
+      chart: gpu-operator
+      version: "v25.3.4"
+      sourceRef:
+        kind: HelmRepository
+        name: nvidia
+        namespace: flux-system
+      interval: 24h
+  install:
+    createNamespace: true
+  upgrade:
+    remediation:
+      remediateLastFailure: true
+  values:
+    driver:
+      enabled: false
+    toolkit:
+      env:
+      - name: CONTAINERD_SOCKET
+        value: /run/k3s/containerd/containerd.sock
+      - name: CONTAINERD_CONFIG
+        value: /var/lib/rancher/k3s/agent/etc/containerd/config.toml
diff --git a/clusters/default/helm/gpu-operator/gpu-operator-repo.yml b/clusters/default/helm/gpu-operator/gpu-operator-repo.yml
new file mode 100644
index 0000000..cce718c
--- /dev/null
+++ b/clusters/default/helm/gpu-operator/gpu-operator-repo.yml
@@ -0,0 +1,9 @@
+---
+apiVersion: source.toolkit.fluxcd.io/v1
+kind: HelmRepository
+metadata:
+  name: nvidia
+  namespace: flux-system
+spec:
+  interval: 24h
+  url: https://helm.ngc.nvidia.com/nvidia