Deploying multiple instances on a single cluster (IBM Concert Operate on OpenShift)

Learn about deploying multiple instances of IBM Concert Operate on a Red Hat OpenShift Container Platform cluster.

Overview

IBM Concert Operate can be deployed multiple times on a single Red Hat OpenShift cluster. This is beneficial if you want to use IBM Concert Operate for multiple separate domains, such as different business units within your organization.

By deploying multiple instances of IBM Concert Operate on a single cluster, the Red Hat OpenShift control plane overhead is reduced and deployment architecture is simplified because there is only one Red Hat OpenShift cluster to manage. A single certificate manager instance can be shared by multiple IBM Concert Operate instances.

Each instance of IBM Concert Operate is deployed in a dedicated namespace, enabling multiple IBM Concert Operate instances to run on the same cluster without conflict.

Restriction:
The following restrictions apply for deployments of multiple instances of IBM Concert Operate on a single cluster:
  • Each instance of IBM Concert Operate must be installed in a dedicated namespace. If you want to deploy two instances of IBM Concert Operate, you cannot use the namespace concert-operate for both instances. You might use, for example, concert-operate and concert-operate2 for the namespaces.
  • The IBM Concert Operate catalog source must not be shared by instances of IBM Concert Operate. The installation procedures guide you through the creation of separate IBM Concert Operate catalog sources in the dedicated namespaces.
  • If you are installing multiple offline instances of IBM Concert Operate and they use the same registry, then you must mirror the images for each instance to different repositories in the registry. You might use, for example, concert-operate and concert-operate2 for the repository names.
  • The IBM Concert Operate operator must be installed in OwnNamespace mode. For more information, see Red Hat OpenShift installation mode.
  • Each new instance of IBM Concert Operate that you deploy on the cluster must be the same version or a more recent version than any already‑installed instances on the cluster. Older versions cannot be deployed into a cluster with a pre-existing newer version. This is because custom resource definitions (CRDs) are cluster-scoped resources, and overwriting them with an older version would break the newer version.
  • Multiple instances of IBM Concert Operate are only supported for deployments on Red Hat OpenShift. Multiple instances of IBM Concert Operate on a single cluster are not supported for deployments of IBM Concert Operate on Linux.
Note: The deployment of other products in the same Red Hat OpenShift cluster as IBM Concert Operate can affect product functionality.

Product Lifecycle

IBM Concert Operate uses a versioning system of X.Y.Z, where X=version, Y=release, and Z=patch. For example, IBM Concert Operate 4.9.1 is version 4, release 9, patch 1.

The deployment of multiple instances in the same cluster is only possible in IBM Concert Operate 4.9.0 and later.

If you deploy multiple instances of IBM Concert Operate on a single cluster, then the difference between the instances can vary by a maximum of one release. This maximum skew must be maintained even during upgrade. When upgrading multiple instances of IBM Concert Operate on a single cluster, you must upgrade the oldest instance first.

For example, IBM Concert Operate 4.9.1 and v4.10.0 can be installed in the same cluster with IBM Concert Operate v4.9.0. However, you can not install v4.11.0 in the same cluster as a v4.9.0 instance, because of a difference of more than one release between the instances.

Storage Requirements

The storage that you configure must be able to handle the load of multiple instances of IBM Concert Operate performantly.

The storage requirements for IBM Concert Operate must be met for each instance individually within the cluster. When an instance is operating under significant load, each additional instance of IBM Concert Operate must still be able to obtain the documented IOPS from the storage provider. This typically requires isolation at the storage layer. For more information, see Storage requirements.

Multiple instances of IBM Concert Operate can share the same storage class.

On a public cloud, a native storage provider is preferred, as stress from one instance is then less likely to affect the performance of other instances.

Hardware Requirements

The Red Hat OpenShift cluster must have sufficient resources to support the combined requirements of all of its IBM Concert Operate instances.

For more information about the hardware requirements of a single instance of IBM Concert Operate, see Hardware requirements. Where there are multiple instances on the same Red Hat OpenShift cluster, the CPU and memory requirements are multiplied by the number of instances.

As part of the installation procedure a prereq checker script validates if there are sufficient resources.