Moving from on-premises Operational Decision Manager to Cloud Pak for Business Automation

In Cloud Pak for Business Automation, the decisions pattern deploys a containerized version of Operational Decision Manager. It provides core business automation capabilities for large amounts of business decisions.

This diagram shows the steps for moving from an on-premises Operational Decision Manager to one in Cloud Pak for Business Automation:

Diagram shows steps.

The instructions for the steps are in the following sections:

1. Assessing your readiness

Start by assessing your readiness to move to Cloud Pak for Business Automation. Make sure that your installation of Operational Decision Manager is version 9.6 or later. For more information, see Upgrade and migration path External link opens a new window or tab.

Moving to a Cloud Pak requires specialized system and cloud administration skills. For more information, see Targeted role-based user archetypes.

These migration instructions apply to production deployments only. Your on-premises database does not have to match the target Cloud Pak database. For example, you can migrate your on-premises Oracle database to a PostgreSQL database.

The following software is required:

Validate your system readiness for hardware and storage capacity on the Red Hat OpenShift® Container Platform (OCP). For more information, see Detailed system requirements External link opens a new window or tab and Storage requirements.

The preparation steps help minimize the downtime for the on-premises move to OCP. If downtime is not an issue, the steps can be done after or concurrently to installing the Cloud Pak.

Database migration scenarios

Scenario 1: Reusing an existing database

If the latency between the Cloud Pak environment and your database is less than 2 milliseconds, you only have to do the Deploying containerized ODM section, and put your database credentials directly in the custom resource (CR) file.

Scenario 2: Using a different database server with the same provider
If the latency between the container and the database is greater than 2 milliseconds, and you must involve your database administrator:
  1. Follow Cleaning Rule Execution Server data .
  2. Your database administrator copies data from the existing database to the new one.
  3. Follow Deploying containerized ODM and put your database credentials directly in the CR file.
Scenario 3: Using a different database server and a different provider
If the latency between the container and the database is greater than 2 milliseconds:
  1. Follow Cleaning Rule Execution Server data and Preparing for the Decision Center data export.
  2. Follow Deploying containerized ODM.
  3. Follow Moving Rule Execution Server data and Moving Decision Center data.

Limitations

The following Operational Decision Manager features are not supported in a containerized environment. Some of these features do not exist in the concurrent release of Operational Decision Manager (see Learn why you should upgrade External link opens a new window or tab):
  • Repackaging Decision Center and Decision Server customized WAR files
  • Classic rule engine and projects, decision validation services, and Enterprise Java™ Bean rule sessions
  • Invocation through Java APIs and generation of Java client projects that use the Java API
  • Samples console
  • Setting the location of the Decision Center Solr cache

Back to top

2. Preparing to move

After assessing your readiness, prepare for the move.

2a. Cleaning Rule Execution Server data
Follow this procedure to clean up your on-premises Rule Execution Server database:
  1. Log in to Rule Execution Server and remove any unused artifacts in the database:
    1. Under Explorer, click Navigator / Libraries > Clean up Libraries.

      Select all the unused libraries and then click Remove.

    2. Under Explorer, click Navigator / Resources > Clean up Resources.

      Select all the unused resources and then click Remove.

  2. It is not possible to migrate the Decision Warehouse traces to another database. If backup is required:
    1. Use the REST API tab of the console.
    2. Use GET /decisiontraces to retrieve the execution traces.

      The traces are returned in JSON format. The resulting JSON file might be very large.

      To reduce the amount of returned traces, set the parameters start, end, offset, and limit accordingly.

    3. Store them in the location of your choice.
2b. Preparing for the Decision Center data export
Follow this procedure to prepare for the export of your on-premises Decision Center database:
  1. Download and decompress the decisioncenter-dbdump.war_.zip file. To access this file, see Decision Center: Database export utility External link opens a new window or tab.
  2. Deploy the decisioncenter-dbdump.war file to your on-premises application server.
  3. On your application server, map the user to do the export to a group that is mapped to the rtsAdministrator role.

    The dbdump web app uses basic authentication. Make sure you map the rtsAdministrator role to an existing user and group during the deployment of the WAR file. For more information, see Deploying the Decision Center WAR files External link opens a new window or tab.

You have two options to use dbdump to export the Decision Center database.

Option 1

Browse to http://<host>:<port>/decisioncenter-dbdump/:

  1. Provide the JNDI name of the data source. The default value is jdbc/ilogDataSource.
  2. Optionally, click the Lookup button to display details about the data source.
  3. Provide the Decision Center schema name. The option Use data source default indicates that the same schema name as the one computed by the Decision Center web app is used to retrieve the data tables.
  4. Do not check Exclude definitions.
  5. Choose one of the following options to generate a compressed file of the Decision Center database:
    • Click Direct download. The file download starts shortly. No file is saved to the server.
    • Click Save on server. Exports the file directly to the server. When the export is complete, a link is generated.

Option 2

Use the following URL to initiate the download: http://<host>:<port>/decisioncenter-dbdump/services/download?<parameters...>.

The <parameters...> correspond to the following arguments:

  • ds: the data source JNDI name (e.g. ds=jdbc/ilogDataSource)
  • schemaName: the name of the schema (e.g. schemaName=DCREPO)
  • useDefaultSchemaName: the request to use the default schema name with useDefaultSchemaName=true

If you encounter a problem, refer to the troubleshooting section in Decision Center: Database export utility External link opens a new window or tab.

Back to top

3. Deploying containerized ODM

After you have assessed your readiness and prepared your data for the move to an OpenShift platform, deploy a Cloud Pak production deployment that includes Operational Decision Manager:

  1. Deploy the operator by using either scripts or the Operator Hub.
  2. Generate the final custom resource (CR) file by using scripts or the templates available in the CASE package.
  3. Modify the generated CR based on your configuration.

    Make sure that the CR has the required database information. You can also migrate your users directly depending on your configuration.

    • Migrate your database.

      See 1. Assessing your readiness for migration scenarios. In some cases, for example when your installations are local, you can reuse your existing database. However, in most production cases, for performance reasons, you need to change the database altogether.

    • Migrating your users.

      For more information about how to set up user access, see Configuring user access.

  4. Deploy the CR to create your instance of Operational Decision Manager.

    See detailed instructions in Deploying the custom resource you created with the deployment script.

Back to top

4. Moving your persisted data

After you have installed and configured the Cloud Pak, you can start moving your data to the new containerized environment.

4a. Moving Rule Execution Server data

Restore your on-premises data into the Decision Server web application in the Cloud Pak.

Clean up the database content as described in Cleaning Rule Execution Server data .
  1. Log in to Rule Execution Server in the on-premises installation.
    1. Navigate to the Server Info tab, and then click Actions > Backup RuleApps.
    2. Save the file.

      The ruleApps.jar file includes all RuleApps, resources, and libraries.

  2. Log in to the Rule Execution Server in the Cloud Pak installation.
    1. Navigate to the Server Info tab, and then click Actions > Restore RuleApps.
    2. Import the ruleApps.jar file that you exported from your on-premises installation of Operational Decision Manager.

      If your ruleApps.jar file is bigger than 500 MB, then the Rule Execution Server console must be tuned to accept such a big backup file. You must also increase the memory that is allocated to the console.

    3. Click Restore.
      Note: All previously deployed RuleApps are deleted.
  3. Migration is complete. Check the Rule Execution Server Explorer tab, and confirm that the RuleApps are the same as in your on-premises installation.
4b. Moving Decision Center data

Move your original on-premises data to the Decision Center of the Cloud Pak.

Generate the Zen API key as described in Generating API keys for authentication External link opens a new window or tab. Then, generate a base64 format of your Zen API key as described in Authorizing HTTP requests by using the Zen API key.

Now that you have a URL to access your Decision Center on the Cloud Pak, you use curl to run an administrative endpoint of the Decision Center REST API that moves the Decision Center data.

The administrative endpoint of the Decision Center REST API does the following tasks:

  • Uploads to the server the database archive that you exported from your on-premises installation, and then expands it into a temporary directory.
  • Imports the expanded archive to your target database in background mode.

You use two Decision Center REST APIs:

  • /v1/repository/dump/import: Imports a database dump into the Decision Center repository. It uses the following parameters:
    • backgroundMode: If set to true, it runs the import as a background task. By default, it is set to false.
    • eraseSchema: If set to true, it deletes and replaces the existing schema. By default, it is set to false.
    • datasource: The data source where the repository is located. By default, it uses jdbc/datasource.
  • /v1/repository/dump/background: Checks the state of the import task that runs in the background. A No response means the import is complete.

For hostport, specify the URL to access Decision Center on the Cloud Pak. Make sure that the URL ends with /decisioncenter-api.

The optional argument <datasource> can be used to specify the data source if it is different from jdbc/ilogDataSource.

To prevent certificate problems when you move your data, use the -k option.

  1. In your on-premises installation of Operational Decision Manager, export your on-premises Decision Center database by using one of the two options that are described in Preparing for the Decision Center data export.
  2. With the base64 format of your Zen API key (XXX), use the following curl command to upload and expand your database archive to the server, and then import it to your new database location: 
    curl -k -H "Authorization: ZenApiKey XXX" -X POST "host:port/odm/decisioncenter-api/v1/repository/dump/import?backgroundMode=true&eraseSchema=true" -H "Content-Type: multipart/form-data" -F "file=@decisioncenter-dbdump.zip;type=application/zip"

    After this command finishes, the import can take some time depending on the size of your database content.

  3. To make sure that the import is complete, run the following command until it returns No, which means that the background task is done: 
    curl -k -H "Authorization: ZenApiKey XXX" -X GET "host:port/odm/decisioncenter-api/v1/repository/dump/background"

  4. When the import is complete, Decision Center in the Cloud Pak is ready to be used.

  5. Review your server definitions in the Decision Center Business Console > Administration > Servers tab. Adjust them with the new URL for Rule Execution Server and Rule Execution Runner.
  6. To verify that everything works as planned, you can run a deployment.

Your original on-premises data has now been moved to the Decision Center of the Cloud Pak.

You can no longer use your on-premises environment. The environments are now separated, and any changes you make on premises are not copied to the cloud.

Back to top

5. Modernizing

After migrating, you can use the new capabilities that are included in Cloud Pak for Business Automation. Modernizing includes updating and improving the architecture of your previous environment.

To customize your migrated data according to your plans, refer to the following configuration parameters:

Back to top