Merge pull request #2367 from DomFleischmann/sync-kubeflow-katib-mani…

…fests-v0.15.0-rc.0 Update kubeflow/katib manifests from v0.15.0-rc.0
kubeflow · Feb 6, 2023 · f7b75c6 · f7b75c6
2 parents a114d58 + c7e301d
commit f7b75c6
Show file tree

Hide file tree

Showing 23 changed files with 335 additions and 50 deletions.
diff --git a/apps/katib/upstream/components/controller/controller.yaml b/apps/katib/upstream/components/controller/controller.yaml
@@ -40,6 +40,17 @@ spec:
             - containerPort: 8080
               name: metrics
               protocol: TCP
+            - containerPort: 18080
+              name: healthz
+              protocol: TCP
+          readinessProbe:
+            httpGet:
+              path: /readyz
+              port: healthz
+          livenessProbe:
+            httpGet:
+              path: /healthz
+              port: healthz
           env:
             - name: KATIB_CORE_NAMESPACE
               valueFrom:

diff --git a/apps/katib/upstream/components/controller/katib-config.yaml b/apps/katib/upstream/components/controller/katib-config.yaml
@@ -8,13 +8,13 @@ data:
   metrics-collector-sidecar: |-
     {
       "StdOut": {
-        "image": "docker.io/kubeflowkatib/file-metrics-collector:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/file-metrics-collector:v0.15.0-rc.0"
       },
       "File": {
-        "image": "docker.io/kubeflowkatib/file-metrics-collector:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/file-metrics-collector:v0.15.0-rc.0"
       },
       "TensorFlowEvent": {
-        "image": "docker.io/kubeflowkatib/tfevent-metrics-collector:v0.14.0-rc.0",
+        "image": "docker.io/kubeflowkatib/tfevent-metrics-collector:v0.15.0-rc.0",
         "resources": {
           "limits": {
             "memory": "1Gi"
@@ -25,42 +25,42 @@ data:
   suggestion: |-
     {
       "random": {
-        "image": "docker.io/kubeflowkatib/suggestion-hyperopt:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/suggestion-hyperopt:v0.15.0-rc.0"
       },
       "tpe": {
-        "image": "docker.io/kubeflowkatib/suggestion-hyperopt:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/suggestion-hyperopt:v0.15.0-rc.0"
       },
       "grid": {
-        "image": "docker.io/kubeflowkatib/suggestion-chocolate:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/suggestion-optuna:v0.15.0-rc.0"
       },
       "hyperband": {
-        "image": "docker.io/kubeflowkatib/suggestion-hyperband:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/suggestion-hyperband:v0.15.0-rc.0"
       },
       "bayesianoptimization": {
-        "image": "docker.io/kubeflowkatib/suggestion-skopt:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/suggestion-skopt:v0.15.0-rc.0"
       },
       "cmaes": {
-        "image": "docker.io/kubeflowkatib/suggestion-goptuna:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/suggestion-goptuna:v0.15.0-rc.0"
       },
       "sobol": {
-        "image": "docker.io/kubeflowkatib/suggestion-goptuna:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/suggestion-goptuna:v0.15.0-rc.0"
       },
       "multivariate-tpe": {
-        "image": "docker.io/kubeflowkatib/suggestion-optuna:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/suggestion-optuna:v0.15.0-rc.0"
       },
       "enas": {
-        "image": "docker.io/kubeflowkatib/suggestion-enas:v0.14.0-rc.0",
+        "image": "docker.io/kubeflowkatib/suggestion-enas:v0.15.0-rc.0",
         "resources": {
           "limits": {
             "memory": "200Mi"
           }
         }
       },
       "darts": {
-        "image": "docker.io/kubeflowkatib/suggestion-darts:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/suggestion-darts:v0.15.0-rc.0"
       },
       "pbt": {
-        "image": "docker.io/kubeflowkatib/suggestion-pbt:v0.14.0-rc.0",
+        "image": "docker.io/kubeflowkatib/suggestion-pbt:v0.15.0-rc.0",
         "persistentVolumeClaimSpec": {
           "accessModes": [
             "ReadWriteMany"
@@ -76,6 +76,6 @@ data:
   early-stopping: |-
     {
       "medianstop": {
-        "image": "docker.io/kubeflowkatib/earlystopping-medianstop:v0.14.0-rc.0"
+        "image": "docker.io/kubeflowkatib/earlystopping-medianstop:v0.15.0-rc.0"
       }
     }
diff --git a/apps/katib/upstream/components/controller/rbac.yaml b/apps/katib/upstream/components/controller/rbac.yaml
@@ -7,39 +7,93 @@ rules:
   - apiGroups:
       - ""
     resources:
-      - configmaps
-      - serviceaccounts
       - services
+    verbs:
+      - "get"
+      - "list"
+      - "watch"
+      - "create"
+      - "delete"
+  - apiGroups:
+      - ""
+    resources:
       - events
-      - namespaces
+    verbs:
+      - "create"
+      - "patch"
+      - "update"
+  - apiGroups:
+      - ""
+    resources:
+      - serviceaccounts
       - persistentvolumes
       - persistentvolumeclaims
+    verbs:
+      - "get"
+      - "list"
+      - "watch"
+      - "create"
+  - apiGroups:
+      - ""
+    resources:
+      - namespaces
+      - configmaps
+    verbs:
+      - "get"
+      - "list"
+      - "watch"
+  - apiGroups:
+      - ""
+    resources:
       - pods
-      - pods/log
       - pods/status
-      - secrets
     verbs:
-      - "*"
+      - "get"
   - apiGroups:
       - apps
     resources:
       - deployments
     verbs:
-      - "*"
+      - "get"
+      - "list"
+      - "watch"
+      - "create"
+      - "delete"
   - apiGroups:
       - rbac.authorization.k8s.io
     resources:
       - roles
       - rolebindings
     verbs:
-      - "*"
+      - "get"
+      - "create"
+      - "list"
+      - "watch"
   - apiGroups:
       - batch
     resources:
       - jobs
       - cronjobs
     verbs:
-      - "*"
+      - "get"
+      - "list"
+      - "watch"
+      - "create"
+      - "delete"
+  - apiGroups:
+      - kubeflow.org
+    resources:
+      - tfjobs
+      - pytorchjobs
+      - mpijobs
+      - xgboostjobs
+      - mxjobs
+    verbs:
+      - "get"
+      - "list"
+      - "watch"
+      - "create"
+      - "delete"
   - apiGroups:
       - kubeflow.org
     resources:
@@ -52,11 +106,6 @@ rules:
       - suggestions
       - suggestions/status
       - suggestions/finalizers
-      - tfjobs
-      - pytorchjobs
-      - mpijobs
-      - xgboostjobs
-      - mxjobs
     verbs:
       - "*"
 ---

diff --git a/apps/katib/upstream/components/controller/service.yaml b/apps/katib/upstream/components/controller/service.yaml
@@ -19,5 +19,8 @@ spec:
     - name: metrics
       port: 8080
       targetPort: 8080
+    - name: healthz
+      port: 18080
+      targetPort: 18080
   selector:
     katib.kubeflow.org/component: controller
diff --git a/apps/katib/upstream/components/controller/trial-templates.yaml b/apps/katib/upstream/components/controller/trial-templates.yaml
@@ -15,7 +15,7 @@ data:
         spec:
           containers:
             - name: training-container
-              image: docker.io/kubeflowkatib/mxnet-mnist:v0.14.0-rc.0
+              image: docker.io/kubeflowkatib/mxnet-mnist:v0.15.0-rc.0
               command:
                 - "python3"
                 - "/opt/mxnet-mnist/mnist.py"
@@ -33,7 +33,7 @@ data:
         spec:
           containers:
             - name: training-container
-              image: docker.io/kubeflowkatib/enas-cnn-cifar10-cpu:v0.14.0-rc.0
+              image: docker.io/kubeflowkatib/enas-cnn-cifar10-cpu:v0.15.0-rc.0
               command:
                 - python3
                 - -u
@@ -54,7 +54,7 @@ data:
             spec:
               containers:
                 - name: pytorch
-                  image: docker.io/kubeflowkatib/pytorch-mnist:v0.14.0-rc.0
+                  image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.15.0-rc.0
                   command:
                     - "python3"
                     - "/opt/pytorch-mnist/mnist.py"
@@ -68,7 +68,7 @@ data:
             spec:
               containers:
                 - name: pytorch
-                  image: docker.io/kubeflowkatib/pytorch-mnist:v0.14.0-rc.0
+                  image: docker.io/kubeflowkatib/pytorch-mnist-cpu:v0.15.0-rc.0
                   command:
                     - "python3"
                     - "/opt/pytorch-mnist/mnist.py"

diff --git a/apps/katib/upstream/components/postgres/kustomization.yaml b/apps/katib/upstream/components/postgres/kustomization.yaml
@@ -0,0 +1,9 @@
+---
+apiVersion: kustomize.config.k8s.io/v1beta1
+kind: Kustomization
+
+resources:
+  - postgres.yaml
+  - pvc.yaml
+  - secret.yaml
+  - service.yaml
diff --git a/apps/katib/upstream/components/postgres/postgres.yaml b/apps/katib/upstream/components/postgres/postgres.yaml
@@ -0,0 +1,42 @@
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: katib-postgres
+  namespace: kubeflow
+  labels:
+    katib.kubeflow.org/component: postgres
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      katib.kubeflow.org/component: postgres
+  strategy:
+    type: Recreate
+  template:
+    metadata:
+      labels:
+        katib.kubeflow.org/component: postgres
+      annotations:
+        sidecar.istio.io/inject: "false"
+    spec:
+      containers:
+        - name: katib-postgres
+          image: postgres:14.5-alpine
+          envFrom:
+            - secretRef:
+                name: katib-postgres-secrets
+          env:
+            - name: PGDATA
+              value: /var/lib/postgresql/data/pgdata
+          ports:
+            - name: postgres
+              containerPort: 5432
+              protocol: TCP
+          volumeMounts:
+            - name: katib-postgres
+              mountPath: /var/lib/postgresql/data
+      volumes:
+        - name: katib-postgres
+          persistentVolumeClaim:
+            claimName: katib-postgres
diff --git a/apps/katib/upstream/components/postgres/pvc.yaml b/apps/katib/upstream/components/postgres/pvc.yaml
@@ -0,0 +1,12 @@
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: katib-postgres
+  namespace: kubeflow
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 3Gi
diff --git a/apps/katib/upstream/components/postgres/secret.yaml b/apps/katib/upstream/components/postgres/secret.yaml
@@ -0,0 +1,10 @@
+---
+apiVersion: v1
+kind: Secret
+type: Opaque
+metadata:
+  name: katib-postgres-secrets
+data:
+  POSTGRES_USER: a2F0aWI=  # katib
+  POSTGRES_PASSWORD: a2F0aWI=  # katib
+  POSTGRES_DB: a2F0aWI=  # katib
diff --git a/apps/katib/upstream/components/postgres/service.yaml b/apps/katib/upstream/components/postgres/service.yaml
@@ -0,0 +1,16 @@
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: katib-postgres
+  namespace: kubeflow
+  labels:
+    katib.kubeflow.org/component: postgres
+spec:
+  type: ClusterIP
+  ports:
+    - port: 5432
+      protocol: TCP
+      name: dbapi
+  selector:
+    katib.kubeflow.org/component: postgres
diff --git a/apps/katib/upstream/components/ui/rbac.yaml b/apps/katib/upstream/components/ui/rbac.yaml
@@ -19,6 +19,18 @@ rules:
       - suggestions
     verbs:
       - "*"
+  - apiGroups:
+      - ""
+    resources:
+      - pods
+    verbs:
+      - list
+  - apiGroups:
+      - ""
+    resources:
+      - pods/log
+    verbs:
+      - get
 ---
 apiVersion: v1
 kind: ServiceAccount

diff --git a/apps/katib/upstream/installs/katib-cert-manager/kustomization.yaml b/apps/katib/upstream/installs/katib-cert-manager/kustomization.yaml
@@ -22,13 +22,13 @@ resources:
 images:
   - name: docker.io/kubeflowkatib/katib-controller
     newName: docker.io/kubeflowkatib/katib-controller
-    newTag: v0.14.0-rc.0
+    newTag: v0.15.0-rc.0
   - name: docker.io/kubeflowkatib/katib-db-manager
     newName: docker.io/kubeflowkatib/katib-db-manager
-    newTag: v0.14.0-rc.0
+    newTag: v0.15.0-rc.0
   - name: docker.io/kubeflowkatib/katib-ui
     newName: docker.io/kubeflowkatib/katib-ui
-    newTag: v0.14.0-rc.0
+    newTag: v0.15.0-rc.0
 
 patchesStrategicMerge:
   - patches/katib-cert-injection.yaml