Deploying multiple IBM Cloud Paks on the same OpenShift cluster

Cloud Pak for Integration 16.1.0 can coexist with other Cloud Paks on the same OpenShift cluster. Use the following procedure to deploy multiple Cloud Paks on the same cluster.

These steps can be carried out in any order. For example, if you already have another Cloud Pak installed on the OpenShift cluster, it is fine to install the Cloud Pak for Integration 16.1.0 operators after that.

Requirements

  • You must install all Cloud Paks in the same installation mode (either in All namespaces on the cluster mode or in A specific namespace on the cluster mode).

  • The optimal deployment approach is to install operators in All namespaces on the cluster mode. For more information, see "Optimal deployment approach" in Structuring your deployment. For more information about installation modes, see Operator reference.
    Important: Not all Cloud Paks support All namespaces on the cluster mode. Ensure the requirements of the other Cloud Paks are understood before choosing a mode, and that all Cloud Paks support the same Foundational services version.

  • When installing in A specific namespace on the cluster mode, the EDB Postgres for Kubernetes operator version must be the same in all namespaces. If the operator is upgraded in one namespace, the operator must be upgraded in all namespaces.

    Downtime and data loss can occur if a Cloud Pak in one namespace downgrades the EDB CustomResourceDefinition (CRD) when an older fixpack is applied, affecting a Cloud Pak in another namespace. For example, if Cloud Pak for Data is installed using EDB 1.18.5, installing Cloud Pak for Integration 16.1.0 will update the CRD to 1.22. If you then then apply a fixpack for Cloud Pak for Data with EDB 1.18.6, the CRD is reverted to the 1.18 version. Fields used by Cloud Pak for Integration in 1.22 that were not in 1.18 will be removed. This can happen both ways as the Cloud Paks move through EDB Postgres versions. This CRD downgrade does not happen when installing in All namespaces on the cluster mode.

Procedure

  1. Install the Cloud Pak for Integration operators. For instructions, see Installing the operators by using the Red Hat OpenShift console or Installing the operators by using the CLI.

  2. Install operators for other Cloud Paks. To get installation procedures for other Cloud Paks, see IBM Cloud Paks documentation.

After completing these steps, you can continue to use Cloud Pak for Integration in the same way that you would for an installation with only Cloud Pak for Integration installed.