Added nodeSelectors for allocating GPU nodePools in the cloud and configured volumes for WebUI

kubernetes/helm/templates/ollama-statefulset.yaml

@@ -19,15 +19,32 @@ spec:
         image: {{ .Values.ollama.image }}
         ports:
         - containerPort: {{ .Values.ollama.servicePort }}
-        resources:
-          limits:
-            cpu: {{ .Values.ollama.resources.limits.cpu }}
-            memory: {{ .Values.ollama.resources.limits.memory }}
-            nvidia.com/gpu: {{ .Values.ollama.resources.limits.gpu }}
+        env:
+        {{- if .Values.ollama.gpu.enabled }}
+          - name: PATH
+            value: /usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
+          - name: LD_LIBRARY_PATH
+            value: /usr/local/nvidia/lib:/usr/local/nvidia/lib64
+          - name: NVIDIA_DRIVER_CAPABILITIES
+            value: compute,utility
+        {{- end}}
+        {{- if .Values.ollama.resources }}
+        resources: {{- toYaml .Values.ollama.resources | nindent 10 }}
+        {{- end }}
         volumeMounts:
         - name: ollama-volume
           mountPath: /root/.ollama
         tty: true
+      {{- with .Values.ollama.nodeSelector }}
+      nodeSelector:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      tolerations:
+        {{- if .Values.ollama.gpu.enabled }}
+        - key: nvidia.com/gpu
+          operator: Exists
+          effect: NoSchedule
+        {{- end }}
   volumeClaimTemplates:
   - metadata:
       name: ollama-volume
@@ -35,4 +52,4 @@ spec:
       accessModes: [ "ReadWriteOnce" ]
       resources:
         requests:
-          storage: 1Gi
+          storage: {{ .Values.ollama.volumeSize }}