[Feat, Misc] Disable PVC creation when pvcStorage is not provided (#176)

* [Add] fix for PVC storage

Signed-off-by: ApostaC <[email protected]>

* [Update] the tutorial of minimal example to exclude PVC configurations

Signed-off-by: ApostaC <[email protected]>

* Bumping the helm chart version

Signed-off-by: ApostaC <[email protected]>

* [fix] condition checking in pvc.yaml

Signed-off-by: ApostaC <[email protected]>

---------

Signed-off-by: ApostaC <[email protected]>

Author: ApostaC

.github/values-01-minimal-example.yaml

@@ -12,10 +12,6 @@ servingEngineSpec:
     requestMemory: "16Gi"
     requestGPU: 1
 
-    pvcStorage: "10Gi"
-    pvcAccessMode:
-      - ReadWriteOnce
-
 routerSpec:
   repository: "localhost:5000/git-act-router"
   imagePullPolicy: "IfNotPresent"

helm/Chart.yaml

@@ -15,7 +15,7 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 0.0.10
+version: 0.0.11
 
 maintainers:
   - name: apostac

helm/templates/deployment-vllm-multi.yaml

@@ -68,7 +68,11 @@ spec:
           imagePullPolicy: IfNotPresent
           env:
           - name: HF_HOME
+            {{- if hasKey $modelSpec "pvcStorage" }}
             value: /data
+            {{- else }}
+            value: /tmp
+            {{- end }}
           {{- with $modelSpec.vllmConfig}}
           {{- if hasKey . "v1" }}
           - name: VLLM_USE_V1
@@ -115,19 +119,23 @@ spec:
               containerPort: {{ include "chart.container-port" . }}
           {{- include "chart.probes" . | indent 10 }}
           resources: {{- include "chart.resources" $modelSpec | nindent 12 }}
+          {{- if hasKey $modelSpec "pvcStorage" }}
           volumeMounts:
           - name: {{ .Release.Name }}-storage
             mountPath: /data
+          {{- end }}
           {{- with $modelSpec.vllmConfig }}
           {{- if hasKey $modelSpec.vllmConfig "tensorParallelSize"}}
           - name: shm
             mountPath: /dev/shm
           {{- end}}
           {{- end}}
       volumes:
+        {{- if hasKey $modelSpec "pvcStorage" }}
         - name: {{ .Release.Name }}-storage
           persistentVolumeClaim:
             claimName: "{{ .Release.Name }}-{{$modelSpec.name}}-storage-claim"
+        {{- end }}
         {{- with $modelSpec.vllmConfig }}
         {{- if hasKey $modelSpec.vllmConfig "tensorParallelSize"}}
         - name: shm

helm/templates/pvc.yaml

@@ -1,5 +1,6 @@
 {{- range $modelSpec := .Values.servingEngineSpec.modelSpec }}
 {{- with $ -}}
+{{- if and (hasKey $modelSpec "pvcStorage") (not (empty $modelSpec.pvcStorage)) }}
 apiVersion: v1
 kind: PersistentVolumeClaim
 metadata:
@@ -24,5 +25,6 @@ spec:
       {{- toYaml $modelSpec.pvcMatchLabels | nindent 8 }}
   {{- end }}
 {{- end }}
+{{- end }}
 ---
 {{- end }}

helm/values.yaml

@@ -21,10 +21,10 @@ servingEngineSpec:
   # - requestMemory: (string) The amount of memory requested for the model, e.g., "16Gi"
   # - requestGPU: (int) The number of GPUs requested for the model, e.g., 1
   #
-  # - pvcStorage: (string) The amount of storage requested for the model, e.g., "50Gi"
-  # - pvcAccessMode: (list) The access mode policy for the mounted volume, e.g., ["ReadWriteOnce"]
-  # - storageClass: (String) The storage class of the PVC e.g., "", default is ""
-  # - pvcMatchLabels: (optional, map) The labels to match the PVC, e.g., {model: "opt125m"}
+  # - pvcStorage: (Optional, string) The amount of storage requested for the model, e.g., "50Gi".
+  # - pvcAccessMode: (Optional, list) The access mode policy for the mounted volume, e.g., ["ReadWriteOnce"]
+  # - storageClass: (Optional, String) The storage class of the PVC e.g., "", default is ""
+  # - pvcMatchLabels: (Optional, map) The labels to match the PVC, e.g., {model: "opt125m"}
   #
   # - vllmConfig: (optional, map) The configuration for the VLLM model, supported options are:
   #   - enablePrefixCaching: (optional, bool) Enable prefix caching, e.g., false

tutorials/01-minimal-helm-installation.md

@@ -49,7 +49,6 @@ servingEngineSpec:
     requestMemory: "16Gi"
     requestGPU: 1
 
-    pvcStorage: "10Gi"
 ```
 
 Explanation of the key fields:
@@ -62,7 +61,6 @@ Explanation of the key fields:
 - **`replicaCount`**: Sets the number of replicas to deploy.
 - **`requestCPU` and `requestMemory`**: Specifies the CPU and memory resource requests for the pod.
 - **`requestGPU`**: Specifies the number of GPUs required.
-- **`pvcStorage`**: Allocates persistent storage for the model.
 
 **Note:** If you intend to set up TWO vllm pods, please refer to `tutorials/assets/values-01-2pods-minimal-example.yaml`.
 

tutorials/assets/values-01-2pods-minimal-example.yaml

@@ -14,10 +14,6 @@ servingEngineSpec:
     requestMemory: "16Gi"
     requestGPU: 0.5
 
-    pvcStorage: "10Gi"
-    pvcAccessMode:
-      - ReadWriteMany
-
     vllmConfig:
       maxModelLen: 1024
       extraArgs: ["--disable-log-requests", "--gpu-memory-utilization", "0.4"]

tutorials/assets/values-01-minimal-example.yaml

@@ -11,7 +11,3 @@ servingEngineSpec:
     requestCPU: 6
     requestMemory: "16Gi"
     requestGPU: 1
-
-    pvcStorage: "10Gi"
-    pvcAccessMode:
-      - ReadWriteOnce