To simplify migration, connection details must be externalized from the transformation (KTR) and job (KJB) files. This allows the same PDI solution to run in different environments (e.g., development, production) by simply changing the external configuration.

Method D involves defining connection parameters (like hostname, username, password) as variables within the central kettle.properties file. PDI automatically loads these variables at startup, making them globally available.

Method B uses a custom properties file for connection details. A parent job reads this file using a step like "Property Input" and then uses the "Set Variables" step to make them available to sub-jobs and transformations. This provides granular control over configuration loading.

Both methods effectively decouple the environment-specific details from the ETL logic.

A. Use the Data Source Wizard to generate the connection details.

This describes a UI tool for creating connections, not the externalization method itself. The connection created could be hard-coded or use variables, so this action alone does not guarantee externalization.

C. Use Manage Data Source to create a new connection in Spoon.

This is another UI action, functionally similar to the Data Source Wizard. It describes how a connection is created in the client, not the underlying strategy for making it portable between environments.

1. Hitachi Vantara Pentaho Documentation

"Use Variables to Streamline Your Processes": This document details the use of variables for parameterizing PDI processes. It explicitly describes the kettle.properties file as the location for defining variables that are global to the PDI environment. This supports using variables in database connections to abstract environment-specific details. (Reference supports option D).

2. Hitachi Vantara Pentaho Documentation

"Set Variables" step reference: The official documentation for the "Set Variables" step explains its function in setting variables dynamically within a job's scope. This is the core component of the pattern described in option B

where a job reads an external properties file and activates those settings for subsequent execution. (Reference supports option B).

3. Hitachi Vantara Pentaho Documentation

"Best Practices - Parameterize Everything": Pentaho best practices strongly advocate for parameterization to support environment migration. They outline the primary methods

including using the kettle.properties file for static environment settings and passing variables/parameters into jobs for more dynamic configurations

which aligns with options D and B.