To install IBM Business Automation Insights, a cluster
administrator user can run a script to set up the cluster. They can also run the script in silent
mode if set of environment variables are created before the script is run. The administrator must
also provide information that they get from the script to a non-administrator user so they can run
the deployment script.
About this task
The cluster setup script is one of four core scripts (cluster setup, prerequisites, deployment,
and post-install) that are provided to help you install IBM Business Automation Insights. You must be a cluster administrator to
run the setup script. For more information, see Cloud Pak personas
.
The cluster setup script identifies or creates a namespace and applies the custom resource
definitions (CRD). It then adds the specified user to the
ibm-bai-insights-engine-operator role, binds the role to the service account.
The script also prompts the administrator to take note of the cluster hostname and a dynamic
storage class on the cluster. These names must be provided to the user who runs the deployment
script.
Note: You can run the scripts on an amd64/x86, a Linux® on Z, or a Linux on Power® machine that can run Podman. The setup script does not set any parameters in the custom resource (CR). The cluster administrator might be running the script on a different host than the user who later runs the deployment script.
A new installation of IBM Business Automation Insights always
includes a namespace-scoped instance of foundational services when you use the scripts.
Use the following steps to complete the setup.
Procedure
-
Download the appropriate repository to a Linux based
machine (RHEL) or a client to a linux-based machine or VM that runs podman natively.
- Optional: If you want to run the script in silent mode, create the
environment variables that are needed for your installation. For more information, see Environment variables for installation in silent
mode
.
-
Log in to the target cluster as the
<cluster-admin> user.
Using the OC client:
oc login https://<cluster-ip>:<port> -u <cluster-admin> -p <password>
On ROKS, if you are not already logged in:
oc login --token=<token> --server=https://<cluster-ip>:<port>
-
Change directory to the extracted cert-kubernetes-bai/scripts
folder.
cd ${PATH_TO_EXTRACTED_FILES}/cert-kubernetes-bai/scripts
- Run the cluster setup script and follow the prompts in the command window.
./bai-clusteradmin-setup.sh
- Select the installation mode: Online or
Offline/Airgap.
- Select the platform type: ROKS (1), OCP (2) or
Other (3).
- If you selected OCP, make sure
Yes is selected if you
want to install Business Automation Insights as a private
catalog rather than in the global catalog namespace (GCN). The GCN uses the
openshift-marketplace namespace, the private option uses the target namespace of
your Business Automation Insights deployment. The default is
Yes.
- Enter Yes or No to confirm whether you want to
install as separation of operators and operands.
- Enter the namespace where you want to install Business Automation Insights.
- Select a non-admin user on your cluster for the installation.
- Enter Yes or No to confirm whether you want to use the images in the IBM® Entitlement Registry.
- If you replied Yes to use the IBM Entitlement Registry, enter your IBM Entitled Registry key and login credentials (user and password).
If you want to load the
container images to a local registry, then set up the cluster by mirroring the images instead of
running the bai-clusteradmin-setup.sh script. For more information, see Setting up the cluster and use a local image registry.
Tip: If you ran the
bai-clusteradmin-setup.sh script and you see one or more of the following
messages, then make sure that you start Docker or Podman and run the script
again.
Error saving credentials: error storing credentials
Error: unable to connect
The Entitlement Registry key failed
- Enter the Fully Qualified Domain Name (FQDN) of the cluster on a Rancher platform. This
is needed for the creation of ingress endpoints.
The following message is displayed:
[INFO] Checking the IBM Cert-manager Operator ready or not
...
[INFO] Applying the latest IBM BAI Operator catalog source...
[✔] IBM BAI Operator catalog source Updated!
-
Monitor the operator pods until they show a STATUS of "Running".
kubectl get pod -w
Tip: If
ibm-bai-insights-engine-operator is inactive for some time,
you can delete the operator pod and let it reconcile.
To confirm that the operator is stuck,
check to see whether the log is providing an output.
kubectl config set-context --current --namespace=<namespace of Business Automation Insights operator
NAMESPACE=<<namespace of Business Automation Insights operator>>
podname=$(kubectl get pod -n $NAMESPACE | grep ibm-bai-insights-engine-operator | awk '{print $1}') kubectl logs $podname -f
You can also list the ClusterServiceVersion (CSV) to verify the version of the
running operators on your cluster.
kubectl get csv -n $NAMESPACE
Note: The version number (25.0.0) of
the installed operators corresponds to the channel for Business Automation Insights
25.0.0.
If you set any subscriptions to manual, then you must approve any pending
operator updates. It is not recommended to set subscriptions to manual because it
can make the installation error prone when some of the dependency operators are not approved. By
default, all subscriptions are set to automatic.
Tip: Subscriptions for the
Cloud Pak foundational services operators are created when they are "needed". Some subscriptions are created during the
installation of the operators. If other subscriptions are needed, they are created during the
installation of the BAI deployment. To check for subscriptions that are waiting for approval, get
the install plans by running the following command.
kubectl get installPlan
Results
When the script is finished, all of the available storage class names are displayed along with
the infrastructure node name. Take a note of the following information and provide it to the admin
user as they are needed for the deployment script:
- Project name or namespace.
- Username to log in to the cluster.
What to do next
You can see the list of operators that are installed in your cluster on the page. For more information about foundational services, see IBM Cloud Pak® foundational services
.
To verify the foundational services installation, check whether all the pods in the target BAI
deployment namespace are running. Use the following command:
kubectl get pods -n $NAMESPACE
Continue to install BAI, you can go to Preparing LDAP secrets for BAI by running a script.