chore: Update ModelMesh helm charts (kubeflow#2315)

These updates coincide with MM 0.9 release. Signed-off-by: Paul Van Eck <pvaneck@us.ibm.com>
magdalenakuhn17 · Jul 6, 2022 · c4d496f · c4d496f
1 parent c4e6b86
commit c4d496f
Show file tree

Hide file tree

Showing 8 changed files with 122 additions and 6 deletions.
diff --git a/charts/kserve/crds/serving.kserve.io_predictor.yaml b/charts/kserve/crds/serving.kserve.io_predictor.yaml
@@ -2,7 +2,7 @@ apiVersion: apiextensions.k8s.io/v1
 kind: CustomResourceDefinition
 metadata:
   annotations:
-    controller-gen.kubebuilder.io/version: v0.6.2
+    controller-gen.kubebuilder.io/version: v0.7.0
   creationTimestamp: null
   labels:
     app.kubernetes.io/instance: modelmesh-controller
@@ -77,6 +77,11 @@ spec:
                 path:
                   description: (DEPRECATED) The path to the model files within the storage
                   type: string
+                protocolVersion:
+                  description:
+                    Protocol version to be exposed by the predictor (i.e.
+                    v1 or v2 or grpc-v1 or grpc-v2)
+                  type: string
                 runtime:
                   description:
                     If omitted a compatible runtime is selected based on
@@ -155,6 +160,7 @@ spec:
                 available: false
                 failedCopies: 0
                 targetModelState: ""
+                totalCopies: 0
                 transitionStatus: UpToDate
               description: PredictorStatus defines the observed state of Predictor
               properties:
@@ -228,6 +234,10 @@ spec:
                     - Loaded
                     - FailedToLoad
                   type: string
+                totalCopies:
+                  default: 0
+                  description: Total number of copies of this predictor's models
+                  type: integer
                 transitionStatus:
                   default: UpToDate
                   description:
@@ -244,6 +254,7 @@ spec:
                 - available
                 - failedCopies
                 - targetModelState
+                - totalCopies
                 - transitionStatus
               type: object
           type: object

diff --git a/charts/kserve/templates/clusterrole.yaml b/charts/kserve/templates/clusterrole.yaml
@@ -338,6 +338,7 @@ rules:
       - ""
     resources:
       - namespaces
+      - namespaces/finalizers
     verbs:
       - get
       - list

diff --git a/charts/kserve/templates/configmap.yaml b/charts/kserve/templates/configmap.yaml
@@ -38,7 +38,7 @@ data:
       limits:
         cpu: "2"
         memory: "512Mi"
-    serviceAccountName: modelmesh
+    serviceAccountName: ""
     metrics:
       enabled: true
 kind: ConfigMap
@@ -203,4 +203,4 @@ data:
 kind: ConfigMap
 metadata:
   name: inferenceservice-config
-  namespace: {{ .Release.Namespace }} 
+  namespace: {{ .Release.Namespace }}
diff --git a/charts/kserve/templates/deployment.yaml b/charts/kserve/templates/deployment.yaml
@@ -164,7 +164,6 @@ spec:
         app.kubernetes.io/managed-by: modelmesh-controller
         app.kubernetes.io/name: modelmesh-controller
         control-plane: modelmesh-controller
-        network-policy: allow-egress
     spec:
       affinity:
         nodeAffinity:

diff --git a/charts/kserve/templates/networkpolicy.yaml b/charts/kserve/templates/networkpolicy.yaml
@@ -0,0 +1,55 @@
+apiVersion: networking.k8s.io/v1
+kind: NetworkPolicy
+metadata:
+  labels:
+    app.kubernetes.io/instance: modelmesh-controller
+    app.kubernetes.io/managed-by: modelmesh-controller
+    app.kubernetes.io/name: modelmesh-controller
+  name: modelmesh-controller
+spec:
+  ingress:
+  - ports:
+    - port: 8443
+      protocol: TCP
+  podSelector:
+    matchLabels:
+      app.kubernetes.io/managed-by: modelmesh-controller
+      control-plane: modelmesh-controller
+  policyTypes:
+  - Ingress
+---
+apiVersion: networking.k8s.io/v1
+kind: NetworkPolicy
+metadata:
+  labels:
+    app.kubernetes.io/instance: modelmesh-controller
+    app.kubernetes.io/managed-by: modelmesh-controller
+    app.kubernetes.io/name: modelmesh-controller
+  name: modelmesh-runtimes
+spec:
+  ingress:
+  - from:
+    - podSelector:
+        matchLabels:
+          app.kubernetes.io/managed-by: modelmesh-controller
+    ports:
+    - port: 8033
+      protocol: TCP
+    - port: 8080
+      protocol: TCP
+  - ports:
+    - port: 8033
+      protocol: TCP
+    - port: 8008
+      protocol: TCP
+  - ports:
+    - port: 2112
+      protocol: TCP
+  podSelector:
+    matchExpressions:
+    - key: modelmesh-service
+      operator: Exists
+    matchLabels:
+      app.kubernetes.io/managed-by: modelmesh-controller
+  policyTypes:
+  - Ingress
diff --git a/charts/kserve/templates/rolebinding.yaml b/charts/kserve/templates/rolebinding.yaml
@@ -5,7 +5,7 @@ metadata:
 roleRef:
   apiGroup: rbac.authorization.k8s.io
   kind: Role
-  name: leader-election-role
+  name: kserve-leader-election-role
 subjects:
 - kind: ServiceAccount
   name: kserve-controller-manager

diff --git a/charts/kserve/templates/servingruntimes.yaml b/charts/kserve/templates/servingruntimes.yaml
@@ -27,6 +27,8 @@ spec:
       value: dummy-model-fixme
     - name: MLSERVER_HOST
       value: 127.0.0.1
+    - name: MLSERVER_GRPC_MAX_MESSAGE_LENGTH
+      value: "-1"
     image: seldonio/mlserver:0.5.2
     name: mlserver
     resources:
@@ -39,6 +41,8 @@ spec:
   grpcDataEndpoint: port:8001
   grpcEndpoint: port:8085
   multiModel: true
+  protocolVersions:
+  - grpc-v2
   supportedModelFormats:
     - name: sklearn
       version: "0"
@@ -52,6 +56,50 @@ spec:
 ---
 apiVersion: serving.kserve.io/v1alpha1
 kind: ServingRuntime
+metadata:
+  labels:
+    app.kubernetes.io/instance: modelmesh-controller
+    app.kubernetes.io/managed-by: modelmesh-controller
+    app.kubernetes.io/name: modelmesh-controller
+    name: modelmesh-serving-ovms-1.x-SR
+  name: ovms-1.x
+spec:
+  builtInAdapter:
+    memBufferBytes: 134217728
+    modelLoadingTimeoutMillis: 90000
+    runtimeManagementPort: 8888
+    serverType: ovms
+  containers:
+  - args:
+    - --port=8001
+    - --rest_port=8888
+    - --config_path=/models/model_config_list.json
+    - --file_system_poll_wait_seconds=0
+    - --grpc_bind_address=127.0.0.1
+    - --rest_bind_address=127.0.0.1
+    image: openvino/model_server:2022.1
+    name: ovms
+    resources:
+      limits:
+        cpu: 5
+        memory: 1Gi
+      requests:
+        cpu: 500m
+        memory: 1Gi
+  grpcDataEndpoint: port:8001
+  grpcEndpoint: port:8085
+  multiModel: true
+  protocolVersions:
+  - grpc-v1
+  supportedModelFormats:
+    - name: openvino_ir
+      version: opset1
+      autoSelect: true
+    - name: onnx
+      version: "1"
+---
+apiVersion: serving.kserve.io/v1alpha1
+kind: ServingRuntime
 metadata:
   annotations:
     maxLoadingConcurrency: "2"
@@ -101,6 +149,8 @@ spec:
   grpcDataEndpoint: port:8001
   grpcEndpoint: port:8085
   multiModel: true
+  protocolVersions:
+  - grpc-v2
   supportedModelFormats:
     - name: keras
       version: "2"

diff --git a/charts/kserve/values.yaml b/charts/kserve/values.yaml
@@ -39,7 +39,7 @@ kserve:
       modelmeshRuntimeAdapterImage: kserve/modelmesh-runtime-adapter
       modelmeshRuntimeAdapterImageTag: *defaultModelMeshVersion
       restProxyImage: kserve/rest-proxy
-      restProxyImageTag: v0.1.1
+      restProxyImageTag: v0.1.4
       podsPerRuntime: 2
   servingruntime:
     modelNamePlaceholder: '{{.Name}}'