Matrix Stack

Added ansible-role-matrix-stack submodule.

.gitmodules

@@ -19,3 +19,6 @@
 [submodule "roles/ericomeehan.libvirt-guest"]
 	path = roles/ericomeehan.libvirt-guest
 	url = git@gitea.eom.dev:DevOps/ansible-role-libvirt-guest.git
+[submodule "roles/ericomeehan.matrix-stack"]
+	path = roles/ericomeehan.matrix-stack
+	url = git@gitea.eom.dev:DevOps/ansible-role-matrix-stack.git

host_vars/alpha-worker-0.yaml

@@ -1 +1,2 @@
 name: alpha-worker-0
+nvidia_driver_needed: true

main.yaml

@@ -64,6 +64,8 @@
     - ../secrets.yaml
   roles:
     - role: ericomeehan.ericomeehan
+    - role: ericomeehan.nvidia_driver
+      when: nvidia_driver_needed == true
 
 - name: Initialize Kubernetes clusters
   hosts: clusters
@@ -153,10 +155,10 @@
         name: bitnami
         repo_url: https://charts.bitnami.com/bitnami
 
-    - name: Add nvdp repository
+    - name: Add nvidia repository
       kubernetes.core.helm_repository:
-        name: nvdp
+        name: nvidia
-        repo_url: https://nvidia.github.io/k8s-device-plugin
+        repo_url: https://helm.ngc.nvidia.com/nvidia
 
     - name: Update Helm repos
       command: helm repo update
@@ -266,12 +268,17 @@
                     ingress:
                       ingressClassName: nginx
 
-    - name: Deploy nvidia device plugin
+    - name: Deploy nvidia gpu operator
       kubernetes.core.helm:
-        name: nvdp
+        name: gpu-operator
-        chart_ref: nvdp/nvidia-device-plugin
+        chart_ref: nvidia/gpu-operator
-        release_namespace: nvidia-device-plugin
+        release_namespace: gpu-operator
         create_namespace: true
+        values:
+          driver:
+            enabled: false
+          toolkit:
+            enabled: false
 
 - name: Port forward HTTP(S) to Ingress Controllers
   hosts: localhost

nvidia-device-plugin.yaml

@@ -4,15 +4,20 @@
   tasks:
     - name: Add nvdp repository
       kubernetes.core.helm_repository:
-        name: nvdp
+        name: nvidia
-        repo_url: https://nvidia.github.io/k8s-device-plugin
+        repo_url: https://helm.ngc.nvidia.com/nvidia
 
     - name: Update Helm repos
       command: helm repo update
 
-    - name: Deploy nvidia device plugin
+    - name: Deploy nvidia gpu operator
       kubernetes.core.helm:
-        name: nvdp
+        name: gpu-operator
-        chart_ref: nvdp/nvidia-device-plugin
+        chart_ref: nvidia/gpu-operator
-        release_namespace: nvidia-device-plugin
+        release_namespace: gpu-operator
         create_namespace: true
+        values:
+          driver:
+            enabled: false
+          toolkit:
+            enabled: false

roles/ericomeehan.matrix-stack

@@ -0,0 +1 @@
+Subproject commit eed66cd1834e81e7eb7c341f4f06d0354ce00f82

test.yaml

@@ -4,4 +4,6 @@
   vars_files:
     - ../secrets.yaml
   roles:
-    - role: ericomeehan.eom
+    - role: ericomeehan.matrix-stack
+      vars:
+        server_name: eom.dev