Preparing and configuring the catalog source for Kubernetes cluster

Edit online

To set up Kubernetes Service cluster and support installation of operators, connect a command line environment to the cluster. The IBM Kubernetes administrator must create the catalog source that registers the IBM Engineering AI Hub operator. After the registration, you can install the operator by using the command line.

About this task

Note: The olm namespace is created and managed by the operator lifecycle manager to run core OLM components such as:

olm-operator
catalog-operator
packageserver

Installing IBM Engineering AI Hub operator-related resources, such as CatalogSource, Subscription, or OperatorGroup into the olm namespace can interfere with OLM functionality, cause installation failures, or lead to unpredictable behavior. Create a namespace to perform the following procedure.

Procedure

Prepare the catalog source manifest. From your local computer, create catalog_source.yaml by using the following content.
```
apiVersion: operators.coreos.com/v1alpha1
    kind: CatalogSource
    metadata:
        name: ibm-engineeringaihub-catalog
        namespace: <catalog_source_namespace>
    spec:
        image: cp.icr.io/cpopen/ibm-elmaihub-operator-catalog:latest
        sourceType: grpc
        updateStrategy:
            registryPoll:
                interval: 10m
```
Note: If you are installing IBM Engineering AI Hub in an air-gapped environment, replace the image name with the local catalog URL.

Note: The updateStrategy must be adjusted to the desired frequency for the operator to check the registry for future updates. In the example, the operator checks for new updates every 10 minutes. To minimize the frequency of checks and reduce resource utilization, you can set the interval to a higher value if required.

Configure the catalog source by using command line.

Run the following command. For <catalog_source_namespace>, enter the name of the catalog source namespace created for the IBM Engineering AI Hub operator. If your catalog source namespace is not available, you can create a new namespace before you apply it.
```
kubectl apply -f catalog_source.yaml -n <catalog_source_namespace> 
```

Verify the installation by running the following command:

kubectl get CatalogSources ibm-engineeringaihub-catalog -n <catalog_source_namespace>

If the installation was successful, the output is similar to this


Name	Display	Type	Publisher	Age
`ibm-engineeringaihub-catalog`	IBM Engineering AI Hub Operator Catalog	grpc	IBM	50s

If the installation fails, see Catalog sources configuration issue for resolution.

The YAML file is imported, and it might take some time for the catalog to connect and reach a ready state.

What to do next

You can start installing the operator on either Red Hat OpenShift cluster or on Kubernetes Service. For more information, see Installing IBM Engineering AI Hub.