System requirements

Before you install the IBM Process Mining, make sure to review the following requirements:

  • Red Hat OpenShift Container Platform 4.7, 4.8, 4.9, 4.10, and 4.11.

  • Persistent storage with the storage class supporting the ReadWriteMany (RWX) mode available on the Red Hat OpenShift cluster.

    Important: The IBM Process Mining Operator automatically provisions the required storage in the IBM Process Mining instance.

    Recommendation: For the production use cases, you pre-create your own persistent storage.

  • MongoDB v4.0 to v5.0 for the production environments.

    Recommendation: For the nonproduction environments, you install the embedded copy of MongoDB Community Edition that is provided in the package.

  • IBM Db2 for the production environments. Supported versions are: 11.5.6.0, 11.5.7.0-cn2

    Recommendation: For the nonproduction environments, you install the embedded copy of IBM Db2 provided in the package.

  • SSL certificates issued by a certificate authority (CA) for the production environments.

    Important: If you did not receive the self-signed certificates for the nonproduction environments, the installation package generates it for you.

Minimum Resources required

Software Memory (GB) CPU (cores) Disk (GB) Nodes
IBM Process Mining 64 16 100 1
IBM Task Mining 32 8 100 1
Total 96 24 200 2

Note

These requirements are related only to the two applications. It means that hardware resources must be allocated also for dependencies (see next paragraph) and OpenShift Container Platform. That is, if you create a cluster with one node (64 Gb RAM and 16core) the resources are shared between IBM Process Mining and Red Hat OpenShift. So the resources available for IBM Process Mining are lower.

Dependencies

The IBM Process Mining Operator depends on the following operators:

  • IBM Cloud Platform Common Services Operator
  • IBM Db2 Operator
  • IBM Redis Operator

Important: If the dependent operators are not present on the cluster, you must install them manually.

Db2 prerequisite

On Red Hat OpenShift Kubernetes Service (ROKS), apply the no root squash command for Db2.

oc get no -l node-role.kubernetes.io/worker --no-headers -o name | xargs -I {} \
   -- oc debug {} \
   -- chroot /host sh -c 'grep "^Domain = slnfsv4.coms" /etc/idmapd.conf || ( sed -i "s/.*Domain =.*/Domain = slnfsv4.com/g" /etc/idmapd.conf; nfsidmap -c; rpc.idmapd )'

Understanding the Operator scope

Operators support the following installation modes:

  • OLM Own namespace Mode
  • OLM All namespace Mode

The IBM Process Mining Operator by default is namespace-scoped. It does not need resources from the cluster. It is recommended to adopt this mode.

It's possible to install the operator in all namespace mode. In this case, you need to be aware that Db2 does not support this capability and you must adopt an external Db2.

By changing WATCH_NAMESPACE variable, it is possible to define namespace watched. If All namespace mode is adopted, it is required to assign dedicated cluster permission.

It is recommended to install only one version of the Operator per cluster.

Understanding the data remediation requirements

You must adhere to the following methods to organize and migrate data accurately in Task Mining:

  • Case 1: Using IBM Task Mining version 1.12.0.3.3 or earlier
    You must install the version IBM Task Mining 1.12.0.3.4 before you migrate to any later versions of the application to ensure successful data migration.
  • Case 2: Using IBM Task Mining version 13.0.0
    You must install the version IBM Task Mining 13.0.1 before you migrate to any later versions of the application. However, you can skip this process if you are directly migrating from 1.12.0.3.4 to later versions.
  • Case 3: Using IBM Task Mining version 1.14.0 or later
    You must upgrade the applications to IBM Task Mining 1.13.2 before you upgrade the applications to IBM Task Mining 1.14.0 or later. However, this requirement is applicable only if the Task Mining application contains projects with RPA snippets that are assigned to tasks.