System requirements for Cloud Pak for Data in an existing IBM Cloud Private installation

Ensure that your servers meet the hardware and software requirements for existing installations of IBM® Cloud Pak for Data Enterprise Edition or Cloud Native Edition.

IBM Cloud Private requirements

To install Cloud Pak for Data to an existing IBM Cloud Private installation, ensure that your IBM Cloud Private installation meets the following minimum requirements:

  • You have IBM Cloud Private Version 3.1.2 or later.

    Only Linux x86_64 installations of IBM Cloud Private are supported.

  • The IBM Cloud Private command-line interface is installed.

    For information, see IBM Cloud Private CLI in the IBM Cloud Private documentation.

  • The user ID that you plan to use to install Cloud Pak for Data has administrator privileges in IBM Cloud Private.
  • You created an image enforcement policy that includes the following entry:
    mycluster.icp:8500/zen/*

    Replace mycluster.icp with your cluster domain name.

    To learn more about creating an image enforcement policy, see "Image security enforcement by using IBM Cloud Private Web console" in Enforcing container image security in the IBM Cloud Private documentation.

  • Your cluster must have one and only one Helm/Tiller instance.

Required resources

Ensure that your IBM Cloud Private cluster has the following resources available:

Available VPCs and memory

The following minimum recommendations are for the Cloud Pak for Data installation.

You must set up the appropriate number of virtual machines or bare metal servers, and choose the specifications that are needed for Cloud Pak for Data. For information about the specifications, see System requirements for stand-alone Cloud Pak for Data installations.

If you plan to install add-ons to Cloud Pak for Data, allocate additional VPCs and memory to Cloud Pak for Data.

Hardware Number of servers (BM/VM) Minimum available VPCs Minimum memory
x86-64 3+ worker nodes.

The nodes must be able to schedule pods.

  • 16 virtual processor cores (VPCs)
  • 64 GB memory

A VPC is a unit of measurement that is used to determine the licensing cost of IBM products. It is based on the number of virtual cores (vCPUs) that are available to the product. A vCPU is a virtual core that is assigned to a virtual machine or a physical processor core if the server isn’t partitioned for virtual machines. A vCPU is equivalent to a Kubernetes CPU. For details, see Meaning of CPU in the Kubernetes documentation.

Requested PersistentVolumeClaim (PVC) size
Ensure that the PVC that you plan use for Cloud Pak for Data has a minimum of 700 GB of storage space.

If you plan to install add-ons to Cloud Pak for Data, allocate additional VPCs and memory to Cloud Pak for Data.

In addition, the PVC must support the ReadWriteMany access mode. For more information, see Access Modes in the Kubernetes documentation.

For information on configuring storage, see Application storage in the IBM Cloud Private documentation.

If you plan to install add-ons on top of Cloud Pak for Data, allocate additional storage space to the PVC.

Docker registry

Ensure that the Docker registry has a minimum of 150 GB of storage space.

If you plan to install add-ons to Cloud Pak for Data, allocate additional storage space to the Docker registry.

By default, Cloud Pak for Data will use mycluster.icp:8500 for the docker registry. Version 2.1.0.1 or earlier: If you plan to use your own Docker registry, or plan to use a namespace within the registry that differs from the namespace where Cloud Pak for Data is installed, then you must create an icpd-configmap.yaml file for the Docker registry with the following contents:

apiVersion: v1
data:
  docker_registry_prefix: docker-registry.default.svc:5000/zen
  docker_registry_secret: sa_zen
metadata:
  labels:
    icpdata-configmap: "true"
  name: icpdata-configmap
  namespace: zen

Where docker-registry.default.svc:5000/zen and sa_zen represent your registry and secret, respectively, where your docker images are loaded. Then create the configmap in the control plane project, for example:

  • In Red Hat OpenShift, enter the command:
    oc create -f icpdata-configmap.yaml
  • In IBM Cloud Private, enter the command:
    kubectl create configmap icpdata-configmap --from-file=icpdata-configmap.yaml

Linux configuration requirements

Log in to each node in the cluster as root and permanently set the Linux kernel semaphore parameter:

echo "kernel.sem = 250 1024000 32 4096" >> /etc/sysctl.conf ;sysctl -p