Put KatibConfig on every install

Signed-off-by: Yuki Iwai <[email protected]>
kubeflow · Jul 31, 2023 · 38de902 · 38de902
1 parent 202a7bb
commit 38de902
Show file tree

Hide file tree

Showing 22 changed files with 424 additions and 121 deletions.
diff --git a/docs/developer-guide.md b/docs/developer-guide.md
@@ -29,7 +29,7 @@ make build REGISTRY=<image-registry> TAG=<image-tag>
 
 To use your custom images for the Katib components, modify
 [Kustomization file](https://github.com/kubeflow/katib/blob/master/manifests/v1beta1/installs/katib-standalone/kustomization.yaml)
-and [Katib Config](https://github.com/kubeflow/katib/blob/master/manifests/v1beta1/components/katib-config/katib-config.yaml)
+and [Katib Config](https://github.com/kubeflow/katib/blob/master/manifests/v1beta1/installs/katib-standalone/katib-config.yaml)
 
 You can deploy Katib v1beta1 manifests into a Kubernetes cluster as follows:
 

diff --git a/docs/new-algorithm-service.md b/docs/new-algorithm-service.md
@@ -90,7 +90,7 @@ Then build the Docker image.
 
 ### Use the algorithm in Katib.
 
-Update the [Katib config](../manifests/v1beta1/components/katib-config/katib-config.yaml) with the new algorithm entity:
+Update the [Katib config](../manifests/v1beta1/installs/katib-standalone/katib-config.yaml) with the new algorithm entity:
 
 ```diff
   runtime:

diff --git a/manifests/v1beta1/components/katib-config/katib-config.yaml b/manifests/v1beta1/components/katib-config/katib-config.yaml
diff --git a/manifests/v1beta1/components/katib-config/kustomization.yaml b/manifests/v1beta1/components/katib-config/kustomization.yaml
diff --git a/manifests/v1beta1/components/katib-config/transformer-configs.yaml b/manifests/v1beta1/components/katib-config/transformer-configs.yaml
diff --git a/manifests/v1beta1/installs/katib-cert-manager/katib-config.yaml b/manifests/v1beta1/installs/katib-cert-manager/katib-config.yaml
@@ -0,0 +1,59 @@
+---
+apiVersion: config.kubeflow.org/v1beta1
+kind: KatibConfig
+init:
+  controller:
+    webhookPort: 8443
+    trialResources:
+      - Job.v1.batch
+      - TFJob.v1.kubeflow.org
+      - PyTorchJob.v1.kubeflow.org
+      - MPIJob.v1.kubeflow.org
+      - XGBoostJob.v1.kubeflow.org
+      - MXJob.v1.kubeflow.org
+runtime:
+  metricsCollectors:
+    - kind: StdOut
+      image: docker.io/kubeflowkatib/file-metrics-collector:latest
+    - kind: File
+      image: docker.io/kubeflowkatib/file-metrics-collector:latest
+    - kind: TensorFlowEvent
+      image: docker.io/kubeflowkatib/tfevent-metrics-collector:latest
+      resources:
+        limits:
+          memory: 1Gi
+  suggestions:
+    - algorithmName: random
+      image: docker.io/kubeflowkatib/suggestion-hyperopt:latest
+    - algorithmName: tpe
+      image: docker.io/kubeflowkatib/suggestion-hyperopt:latest
+    - algorithmName: grid
+      image: docker.io/kubeflowkatib/suggestion-optuna:latest
+    - algorithmName: hyperband
+      image: docker.io/kubeflowkatib/suggestion-hyperband:latest
+    - algorithmName: bayesianoptimization
+      image: docker.io/kubeflowkatib/suggestion-skopt:latest
+    - algorithmName: cmaes
+      image: docker.io/kubeflowkatib/suggestion-goptuna:latest
+    - algorithmName: sobol
+      image: docker.io/kubeflowkatib/suggestion-goptuna:latest
+    - algorithmName: multivariate-tpe
+      image: docker.io/kubeflowkatib/suggestion-optuna:latest
+    - algorithmName: enas
+      image: docker.io/kubeflowkatib/suggestion-enas:latest
+      resources:
+        limits:
+          memory: 200Mi
+    - algorithmName: darts
+      image: docker.io/kubeflowkatib/suggestion-darts:latest
+    - algorithmName: pbt
+      image: docker.io/kubeflowkatib/suggestion-pbt:latest
+      persistentVolumeClaimSpec:
+        accessModes:
+          - ReadWriteMany
+        resources:
+          requests:
+            storage: 5Gi
+  earlyStoppings:
+    - algorithmName: medianstop
+      image: docker.io/kubeflowkatib/earlystopping-medianstop:latest
diff --git a/manifests/v1beta1/installs/katib-cert-manager/kustomization.yaml b/manifests/v1beta1/installs/katib-cert-manager/kustomization.yaml
@@ -19,8 +19,6 @@ resources:
   - ../../components/webhook/
   # Cert-manager certificate for webhooks
   - certificate.yaml
-  # Katib Config.
-  - ../../components/katib-config/
 images:
   - name: docker.io/kubeflowkatib/katib-controller
     newName: docker.io/kubeflowkatib/katib-controller
@@ -61,3 +59,11 @@ vars:
 
 configurations:
   - params.yaml
+
+configMapGenerator:
+  - name: katib-config
+    behavior: create
+    files:
+      - katib-config.yaml
+    options:
+      disableNameSuffixHash: true
diff --git a/manifests/v1beta1/installs/katib-external-db/katib-config.yaml b/manifests/v1beta1/installs/katib-external-db/katib-config.yaml
@@ -0,0 +1,59 @@
+---
+apiVersion: config.kubeflow.org/v1beta1
+kind: KatibConfig
+init:
+  controller:
+    webhookPort: 8443
+    trialResources:
+      - Job.v1.batch
+      - TFJob.v1.kubeflow.org
+      - PyTorchJob.v1.kubeflow.org
+      - MPIJob.v1.kubeflow.org
+      - XGBoostJob.v1.kubeflow.org
+      - MXJob.v1.kubeflow.org
+runtime:
+  metricsCollectors:
+    - kind: StdOut
+      image: docker.io/kubeflowkatib/file-metrics-collector:latest
+    - kind: File
+      image: docker.io/kubeflowkatib/file-metrics-collector:latest
+    - kind: TensorFlowEvent
+      image: docker.io/kubeflowkatib/tfevent-metrics-collector:latest
+      resources:
+        limits:
+          memory: 1Gi
+  suggestions:
+    - algorithmName: random
+      image: docker.io/kubeflowkatib/suggestion-hyperopt:latest
+    - algorithmName: tpe
+      image: docker.io/kubeflowkatib/suggestion-hyperopt:latest
+    - algorithmName: grid
+      image: docker.io/kubeflowkatib/suggestion-optuna:latest
+    - algorithmName: hyperband
+      image: docker.io/kubeflowkatib/suggestion-hyperband:latest
+    - algorithmName: bayesianoptimization
+      image: docker.io/kubeflowkatib/suggestion-skopt:latest
+    - algorithmName: cmaes
+      image: docker.io/kubeflowkatib/suggestion-goptuna:latest
+    - algorithmName: sobol
+      image: docker.io/kubeflowkatib/suggestion-goptuna:latest
+    - algorithmName: multivariate-tpe
+      image: docker.io/kubeflowkatib/suggestion-optuna:latest
+    - algorithmName: enas
+      image: docker.io/kubeflowkatib/suggestion-enas:latest
+      resources:
+        limits:
+          memory: 200Mi
+    - algorithmName: darts
+      image: docker.io/kubeflowkatib/suggestion-darts:latest
+    - algorithmName: pbt
+      image: docker.io/kubeflowkatib/suggestion-pbt:latest
+      persistentVolumeClaimSpec:
+        accessModes:
+          - ReadWriteMany
+        resources:
+          requests:
+            storage: 5Gi
+  earlyStoppings:
+    - algorithmName: medianstop
+      image: docker.io/kubeflowkatib/earlystopping-medianstop:latest
diff --git a/manifests/v1beta1/installs/katib-external-db/kustomization.yaml b/manifests/v1beta1/installs/katib-external-db/kustomization.yaml
@@ -17,8 +17,6 @@ resources:
   - ../../components/cert-generator/
   # Katib webhooks.
   - ../../components/webhook/
-  # Katib Config.
-  - ../../components/katib-config/
 images:
   - name: docker.io/kubeflowkatib/katib-controller
     newName: docker.io/kubeflowkatib/katib-controller
@@ -39,3 +37,10 @@ secretGenerator:
   - name: katib-mysql-secrets
     envs:
       - secrets.env
+configMapGenerator:
+  - name: katib-config
+    behavior: create
+    files:
+      - katib-config.yaml
+    options:
+      disableNameSuffixHash: true
diff --git a/manifests/v1beta1/installs/katib-leader-election/katib-config.yaml b/manifests/v1beta1/installs/katib-leader-election/katib-config.yaml
@@ -1,5 +1,3 @@
-# This KatibConfig is mostly same as https://github.com/kubeflow/katib/manifests/v1beta1/components/katib-config/katib-config.yaml.
-# Only `.init.controller.enableLeaderElection` field is different.
 ---
 apiVersion: config.kubeflow.org/v1beta1
 kind: KatibConfig
@@ -17,40 +15,40 @@ init:
 runtime:
   metricsCollectors:
     - kind: StdOut
-      image: docker.io/kubeflowkatib/file-metrics-collector:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/file-metrics-collector:latest
     - kind: File
-      image: docker.io/kubeflowkatib/file-metrics-collector:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/file-metrics-collector:latest
     - kind: TensorFlowEvent
-      image: docker.io/kubeflowkatib/tfevent-metrics-collector:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/tfevent-metrics-collector:latest
       resources:
         limits:
           memory: 1Gi
   suggestions:
     - algorithmName: random
-      image: docker.io/kubeflowkatib/suggestion-hyperopt:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-hyperopt:latest
     - algorithmName: tpe
-      image: docker.io/kubeflowkatib/suggestion-hyperopt:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-hyperopt:latest
     - algorithmName: grid
-      image: docker.io/kubeflowkatib/suggestion-optuna:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-optuna:latest
     - algorithmName: hyperband
-      image: docker.io/kubeflowkatib/suggestion-hyperband:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-hyperband:latest
     - algorithmName: bayesianoptimization
-      image: docker.io/kubeflowkatib/suggestion-skopt:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-skopt:latest
     - algorithmName: cmaes
-      image: docker.io/kubeflowkatib/suggestion-goptuna:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-goptuna:latest
     - algorithmName: sobol
-      image: docker.io/kubeflowkatib/suggestion-goptuna:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-goptuna:latest
     - algorithmName: multivariate-tpe
-      image: docker.io/kubeflowkatib/suggestion-optuna:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-optuna:latest
     - algorithmName: enas
-      image: docker.io/kubeflowkatib/suggestion-enas:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-enas:latest
       resources:
         limits:
           memory: 200Mi
     - algorithmName: darts
-      image: docker.io/kubeflowkatib/suggestion-darts:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-darts:latest
     - algorithmName: pbt
-      image: docker.io/kubeflowkatib/suggestion-pbt:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/suggestion-pbt:latest
       persistentVolumeClaimSpec:
         accessModes:
           - ReadWriteMany
@@ -59,4 +57,4 @@ runtime:
             storage: 5Gi
   earlyStoppings:
     - algorithmName: medianstop
-      image: docker.io/kubeflowkatib/earlystopping-medianstop:$(KATIB_VERSION)
+      image: docker.io/kubeflowkatib/earlystopping-medianstop:latest
diff --git a/manifests/v1beta1/installs/katib-openshift/katib-config.yaml b/manifests/v1beta1/installs/katib-openshift/katib-config.yaml
@@ -0,0 +1,59 @@
+---
+apiVersion: config.kubeflow.org/v1beta1
+kind: KatibConfig
+init:
+  controller:
+    webhookPort: 8443
+    trialResources:
+      - Job.v1.batch
+      - TFJob.v1.kubeflow.org
+      - PyTorchJob.v1.kubeflow.org
+      - MPIJob.v1.kubeflow.org
+      - XGBoostJob.v1.kubeflow.org
+      - MXJob.v1.kubeflow.org
+runtime:
+  metricsCollectors:
+    - kind: StdOut
+      image: docker.io/kubeflowkatib/file-metrics-collector:latest
+    - kind: File
+      image: docker.io/kubeflowkatib/file-metrics-collector:latest
+    - kind: TensorFlowEvent
+      image: docker.io/kubeflowkatib/tfevent-metrics-collector:latest
+      resources:
+        limits:
+          memory: 1Gi
+  suggestions:
+    - algorithmName: random
+      image: docker.io/kubeflowkatib/suggestion-hyperopt:latest
+    - algorithmName: tpe
+      image: docker.io/kubeflowkatib/suggestion-hyperopt:latest
+    - algorithmName: grid
+      image: docker.io/kubeflowkatib/suggestion-optuna:latest
+    - algorithmName: hyperband
+      image: docker.io/kubeflowkatib/suggestion-hyperband:latest
+    - algorithmName: bayesianoptimization
+      image: docker.io/kubeflowkatib/suggestion-skopt:latest
+    - algorithmName: cmaes
+      image: docker.io/kubeflowkatib/suggestion-goptuna:latest
+    - algorithmName: sobol
+      image: docker.io/kubeflowkatib/suggestion-goptuna:latest
+    - algorithmName: multivariate-tpe
+      image: docker.io/kubeflowkatib/suggestion-optuna:latest
+    - algorithmName: enas
+      image: docker.io/kubeflowkatib/suggestion-enas:latest
+      resources:
+        limits:
+          memory: 200Mi
+    - algorithmName: darts
+      image: docker.io/kubeflowkatib/suggestion-darts:latest
+    - algorithmName: pbt
+      image: docker.io/kubeflowkatib/suggestion-pbt:latest
+      persistentVolumeClaimSpec:
+        accessModes:
+          - ReadWriteMany
+        resources:
+          requests:
+            storage: 5Gi
+  earlyStoppings:
+    - algorithmName: medianstop
+      image: docker.io/kubeflowkatib/earlystopping-medianstop:latest