Chapter 3. Preparing OpenShift AI for use in IBM Cloud Pak for Data

Procedure

1. Open a new terminal window.

2. In the OpenShift command-line interface (CLI), log in to your OpenShift Container Platform cluster as a cluster administrator, as shown in the following example:

   $ oc login <openshift_cluster_url> -u <admin_username> -p <password>

   Copy to Clipboard Toggle word wrap

3. Create a new YAML file with the following contents:

   apiVersion: v1
   kind: Namespace 

   1

   
   metadata:
     name: redhat-ods-operator
   ---
   apiVersion: operators.coreos.com/v1
   kind: OperatorGroup 

   2

   
   metadata:
     name: rhods-operator
     namespace: redhat-ods-operator
   ---
   apiVersion: operators.coreos.com/v1alpha1
   kind: Subscription 

   3

   
   metadata:
     name: rhods-operator
     namespace: redhat-ods-operator
   spec:
     name: rhods-operator
     channel: eus-2.8  

   4

   
     source: redhat-operators
     sourceNamespace: openshift-marketplace
     config:
        env:
           - name: "DISABLE_DSC_CONFIG"

   Copy to Clipboard Toggle word wrap

   1

   Defines the recommended redhat-ods-operator namespace for installation of the Operator.

   2

   Defines an Operator group for installation of the Operator. You must specify the same namespace that you defined earlier in the file. Also, another Operator group must not exist in the namespace.

   3

   Defines a subscription for the Operator. You must specify the same namespace that you defined earlier in the file.

   4

   For channel, specify a value of eus-2.8. This channel name indicates that OpenShift AI 2.8 is an Extended Update Support (EUS) version. For more information, see Red Hat OpenShift AI Self-Managed Life Cycle.

4. Deploy the YAML file to create the namespace, Operator group, and subscription that you defined.

   $ oc create -f <file_name>.yaml

   Copy to Clipboard Toggle word wrap

   You see output similar to the following:

   namespace/redhat-ods-operator created
   operatorgroup.operators.coreos.com/rhods-operator created
   subscription.operators.coreos.com/rhods-operator created

   Copy to Clipboard Toggle word wrap

5. Create another new YAML file with the following contents:

   apiVersion: dscinitialization.opendatahub.io/v1
   kind: DSCInitialization 

   1

   
   metadata:
     name: default-dsci
   spec:
     applicationsNamespace: redhat-ods-applications
     monitoring:
       managementState: Managed
       namespace: redhat-ods-monitoring
     serviceMesh:
       managementState: Removed 

   2

   
     trustedCABundle:
       managementState: Managed
       customCABundle: ""

   Copy to Clipboard Toggle word wrap

   1

   Defines a DSCInitialization object called default-dsci. The DSCInitialization object is used by the Operator to manage resources that are common to all product components.

   2

   Specifies that the serviceMesh component (which is a dependency for KServe in some setups) is not installed. This setting is required when preparing OpenShift AI for use in IBM products.

Procedure

1. Open a new terminal window.

2. In the OpenShift command-line interface (CLI), log in to your OpenShift Container Platform cluster as a cluster administrator, as shown in the following example:

   $ oc login <openshift_cluster_url> -u <admin_username> -p <password>

   Copy to Clipboard Toggle word wrap

3. Create a new YAML file with the following contents:

   kind: DataScienceCluster 

   1

   
   apiVersion: datasciencecluster.opendatahub.io/v1
   metadata:
     name: default-dsc
     labels:
       app.kubernetes.io/name: datasciencecluster
       app.kubernetes.io/instance: default-dsc
       app.kubernetes.io/part-of: rhods-operator
       app.kubernetes.io/managed-by: kustomize
       app.kubernetes.io/created-by: rhods-operator
   spec:
     components:
       codeflare:
         managementState: Removed
       dashboard:
         managementState: Removed
       datasciencepipelines:
         managementState: Removed
       kserve:
         managementState: Managed
         defaultDeploymentMode: RawDeployment 

   2

   
         serving:
           managementState: Removed 

   3

   
           name: knative-serving
       kueue:
         managementState: Removed
       modelmeshserving:
         managementState: Removed
       ray:
         managementState: Removed
       workbenches:
         managementState: Removed

   Copy to Clipboard Toggle word wrap

   1

   Defines a new DataScienceCluster object called default-dsc. The DataScienceCluster object is used by the Operator to manage OpenShift AI components.

   2

   Specifies that the KServe component is installed and managed by the Operator in raw mode, which means that KServe does not have any other components as dependencies.

   3

   Specifies that the serving component (which is a dependency for KServe in some setups) is not installed. This setting is required when preparing OpenShift AI for use in Cloud Pak for Data.

   In addition, observe that the value of the managementState field for all other OpenShift AI components is set to Removed. This means that the components are not installed.

   In general, the Managed and Removed values are described as follows:

   Managed

   The Operator actively manages the component, installs it, and tries to keep it active. The Operator will upgrade the component only if it is safe to do so.

   Removed

   The Operator actively manages the component but does not install it. If the component is already installed, the Operator will try to remove it.

4. Create the DataScienceCluster object on your OpenShift cluster.

   $ oc create -f <file_name>.yaml

   Copy to Clipboard Toggle word wrap

   You see output similar to the following:

   datasciencecluster.datasciencecluster.opendatahub.io/default-dsc created

   Copy to Clipboard Toggle word wrap

Procedure

1. Log in to the OpenShift Container Platform web console as a cluster administrator.
2. In the web console, click Workloads ConfigMaps.
3. In the Project list, click redhat-ods-applications.
4. On the list of ConfigMap resources, click the inferenceservice-config resource and then click the YAML tab.

5. In the metadata.annotations section of the file, add opendatahub.io/managed: 'false' as shown:

   metadata:
     annotations:
       internal.config.kubernetes.io/previousKinds: ConfigMap
       internal.config.kubernetes.io/previousNames: inferenceservice-config
       internal.config.kubernetes.io/previousNamespaces: opendatahub
       opendatahub.io/managed: 'false'

   Copy to Clipboard Toggle word wrap

   The additional annotation makes the inferencing configuration file editable.

6. To prevent errors when using OpenShift AI in Cloud Pak for Data (including watsonx.ai), update the configuration as follows:

   1. In the YAML file, locate the following entry:

      "domainTemplate": "{{ .Name }}-{{ .Namespace }}.{{ .IngressDomain }}

      Copy to Clipboard Toggle word wrap

   2. Update the value of the domainTemplate field as shown:

      "domainTemplate": "example.com"

      Copy to Clipboard Toggle word wrap

      This new, explicitly specified value ensures that OpenShift AI cannot generate a value for the domainTemplate field that exceeds the maximum length and causes an error. The domainTemplate field is not used by the raw deployment mode that you configured for the KServe component when preparing OpenShift AI for use in Cloud Pak for Data.

7. To directly mount the volume and avoid copying the model, update the value of the enableDirectPvcVolumeMount field as shown:

   "enableDirectPvcVolumeMount": true

   Copy to Clipboard Toggle word wrap
8. Click Save.