Installing operators
IBM Cloud Pak® for Integration operators are installed and managed using the Operator Lifecycle Manager (OLM) in Red Hat OpenShift. After the operators are installed, you can use them to deploy Cloud Pak for Integration capabilities and runtimes.
Operators extend a Kubernetes cluster by adding and managing additional resource types in the Kubernetes API. This enables the installation and management of software using standard Kubernetes tools. To learn more, see Understanding Operators in the Red Hat OpenShift documentation.
Operators available to install
You can install any combination of operators. Any dependencies will be pulled in automatically.
Each operator below is listed by its name in the OpenShift web console UI, followed by its literal that is used in the CLI.
- IBM Cloud Pak for Integration
Top level Cloud Pak for Integration operator that will install all other Cloud Pak for Integration operators automatically. Use this to install the whole Cloud Pak in one operation.
name:ibm-cp-integration
- IBM Cloud Pak for Integration Platform Navigator
Provides a dashboard and central services for other Cloud Pak for Integration capabilities. Should be installed for most Cloud Pak for Integration installations.
name:ibm-integration-platform-navigator
- IBM Automation Foundation assets
Stores, manages, retrieves and searches for integration assets for use within the IBM Cloud Pak for Integration and its capabilities.
name:ibm-integration-asset-repository
- IBM Cloud Pak for Integration Operations Dashboard
Transaction tracing across capabilities and runtimes to allow troubleshooting and investigation of errors and latency issues across integration capabilities to ensure applications meet service level agreements.
name:ibm-integration-operations-dashboard
- IBM API Connect
Provides the API management and Event Endpoint Management capabilities.
name:ibm-apiconnect
- IBM App Connect
Provides application integration capabilities and a means to easily create and export flows that run in an App Connect instance.
name:ibm-appconnect
- IBM Aspera HSTS
Provides high speed transfer integration capabilities.
name:aspera-hsts-operator
- IBM DataPower Gateway
Provides gateway capabilities.
name:datapower-operator
- IBM MQ
Provides messaging capabilities.
name:ibm-mq
Installation methods
There are two methods for installing operators:
OpenShift web console
oc
command-line tool (CLI)
OpenShift web console
Log into the OpenShift web console with your OpenShift cluster admin credentials.
Make sure you have selected the Administrator view.
Click Operators > OperatorHub > Integration & Delivery.
Click the tile for the operator you want to install, for example, IBM Cloud Pak for Integration.
Click Install.
In the Install Operator pane:
Select the desired operator channel. See Operator channel versions for IBM Cloud Pak for Integration releases for more information.
Select the option to install Cloud Pak for Integration in one namespace or in all namespaces in your cluster. See "Guidelines for installing operators" below for more information.
Select an approval strategy. You can select the automatic or manual option. For more information about a manual approval strategy, see Restricting automatic updates with an approval strategy.
Click Install.
oc
command-line tool
Only if your desired installation mode is a specific namespace on the cluster, create an
OperatorGroup
similar to the example below. Replace<namespace>
with your desired namespace.apiVersion: operators.coreos.com/v1 kind: OperatorGroup metadata: name: ibm-integration-operatorgroup namespace: <namespace> spec: targetNamespaces: - <namespace>
Create a
Subscription
similar to the example below for the IBM Cloud Pak for Integration operator. Replace<namespace>
with:openshift-operators
, if installing the operator in All namespaces on the cluster.Your chosen namespace, if installing in A specific namespace on the cluster.
apiVersion: operators.coreos.com/v1alpha1 kind: Subscription metadata: name: ibm-cp-integration namespace: <namespace> spec: channel: v1.3 name: ibm-cp-integration source: ibm-operator-catalog sourceNamespace: openshift-marketplace
Guidelines for installing operators
For additional details and deployment recommendations, see Structuring your deployment.
default
, kube-system
, kube-public
,openshift-node
, openshift-infra
, and openshift
.When installing Cloud Pak for Integration operators, follow these guidelines:
Typically, a cluster administrator installs the operators, and an automation administrator creates the customer resources (installed instance of the operator) and configures the capabilities and runtimes. See Roles for more information.
Operators can be installed at cluster scope (in all namespaces in the cluster) or at namespace scope (in an individual namespace).
If the operators are installed at cluster scope, the entire cluster effectively behaves as one large tenant.
If the operators are installed at namespace scope, each namespace effectively behaves as a different tenant.
Cloud Pak for Integration can be installed in either of two installation modes: All namespaces on the cluster or A specific namespace on the cluster.
All IBM Cloud Paks installed in a cluster must be installed in the same mode, whether All namespaces on the cluster or A specific namespace on the cluster. If you are installing one of more Cloud Paks with Cloud Pak for Integration, you can't use both mode types together.
When Cloud Pak for Integration is installed in the All namespaces on the cluster installation mode, there can be only one Platform Navigator per cluster, and all Cloud Pak instances are owned by that Platform Navigator.
When Cloud Pak for Integrationis installed in A specific namespace on the cluster mode, there can be one Platform Navigator in each namespace, and that Platform Navigator owns only the instances in that namespace.
When operators are installed in the All namespaces on the cluster installation mode, the operator must be installed in the openshift-operators namespace.
For both installation modes, by default a single instance of IBM Cloud Pak foundational services is installed in the
ibm-common-services
namespace if the foundational services operator is not already installed on the cluster. The foundational services operator must be installed in the same mode as the Cloud Pak operators. However, regardless of the mode with which it is installed, by default the foundational services operator will install the Operand Deployment Lifecycle Manager operator in the A specific namespace on the cluster mode in theibm-common-services
namespace. While the Operand Deployment Lifecycle Manager operator is installed in this mode, it will extend its permissions to manage foundational services resources in all the namespaces where capabilities and runtimes of the Cloud Pak are installed. To use a custom namespace for foundational services, see Installing in a custom namespace before installing Cloud Pak for Integration operators.When Cloud Pak for Integration is installed in either mode, operators still need a small set of cluster level permissions to allow manipulation of resources defined at cluster scope, such as reading Custom Resource Definitions. See Cluster-scoped permissions for information about the permissions required.
If you need the entire Cloud Pak for Integration (or only certain of its capabilities and runtimes) to use manual approval—and not update automatically—see Restricting automatic updates with an approval strategy.
Validating an installed operator
OpenShift web console
Log into the OpenShift web console.
In the navigation menu, click Operators > Installed Operators.
Select the project in which the operator was installed. If you installed the operator using the All namespaces on the cluster installation mode, the operator pod is located in the
openshift-operators
project.Locate your Operator in the table and examine the
status
column. When the Operator is ready, it will have the messageSucceeded
.
oc
command-line tool
To check the progress of operator deployment, you can look at the operator pod within the selected namespace using the oc
command-line tool.
You can now proceed to install other Cloud Pak for Integration operators.
Restricting automatic updates with an approval strategy
Operators can be upgraded automatically when new compatible versions are available. However, in some situations this automatic upgrade may be undesirable. You can control whether or not an operator is upgraded automatically by setting an approval strategy, either:
When the operator is installed.
After installation, by modifying its subscription.
The available approval strategies are:
Automatic (default): As new operator versions are made available on the subscription channel, they are installed automatically.
Manual: As new operator versions are made available on the subscription channel, the subscription indicates that there is an update available, but you have to approve the update manually.
Using the OpenShift console
To confirm or update the approval strategy for an operator in the OpenShift console:
Expand the Operators section in the navigation bar on the left, and select Installed Operators. Ensure you select the project containing the operators you intend to confirm or modify.
Click the operator you want to confirm or modify.
Click the Subscription tab.
The Subscription Details at the top of the tab will show you the Channel referenced by the subscription, what the Approval strategy is, and what the Upgrade Status is.
If the Approval strategy is set to Automatic, the Upgrade Status should show Up to date.
If the Approval strategy is set to Manual, the Upgrade Status may show Upgrade available. In this case, there may be an available control that allows you to approve the InstallPlan that upgrades the operator.
You can also change the approval strategy by clicking the Subscription tab, then selecting the Approval strategy value. A Change Update Approval Strategy dialog appears and allows you to select a different strategy.
Using the OpenShift CLI
If you have both the OpenShift CLI with oc
and the jq JSON stream editor installed, you can confirm or update the approval strategy for all subscriptions in a specific project/namespace. To do so, use the oc
command in a bash
shell:
NAMESPACE=<your namespace>
subscriptions=$(oc get subscription -n $NAMESPACE -o name)
for subscription in ${subscriptions}; do
echo -n "${subscription} "
oc get ${subscription} -n $NAMESPACE -o json | jq '.spec.installPlanApproval'
done
If you have IBM operators installed across different namespaces, you can obtain a list of all subscriptions and their namespaces with:
oc get subscription --all-namespaces -o custom-columns=NAME:.metadata.name,NAMESPACE:.metadata.namespace \
| grep 'ibm-' \
| xargs -n2 sh -c 'printf "$0 $1: " && oc get subscription "$0" -n "$1" -o json | jq ".spec.installPlanApproval"'
This returns a row for each subscription, containing the subscription name, its namespace, and the approval strategy. For example:
ibm-cert-manager-operator ibm-common-services: "Automatic"
ibm-common-service-operator ibm-common-services: "Automatic"
ibm-commonui-operator ibm-common-services: "Automatic"
...
ibm-cp-integration cp4i: "Manual"
...
IBM Cloud Pak foundational services approval strategies
The approval strategy for foundational services is managed using the CommonService
configuration resource. By default, foundational services is installed in a separate namespace, which means that its approval strategy must be set separately. However, if you choose to install foundational services in the same namespace as your other operators, it inherits the approval strategy selected for your other operators.
For more information, see Approval strategy in "Configuring IBM Cloud Pak foundational services by using the CommonService custom resource".