madconfig utility usage for virtual MDM implementations

The madconfig utility is internally used by IBM® InfoSphere® Master Data Management components and is used externally to configure various components.

This utility is run from the operational server installation MDM_INSTALL_HOME/MDS/scripts directory.

The targets described in this section are applicable for use with a virtual MDM implementation. In addition to the targets listed here, madconfig is also used in the installation process. You can read about install-specific targets in that topic.

Before you run a utility, make sure that you have set the necessary operational server environment variables. For information about the variables, see the operational server environment variables topic.


Although you can use the madconfig utility to generate and validates weights, it is suggested that you do these tasks from InfoSphere MDM Workbench.

Table 1. madconfig utility options
Options and targets Description
-projecthelp Lists the valid options.
-propertyfile Loads properties from file.
-recordfile Record response properties to file. For details on using the -recordfile option, see the Creating an automated madconfig utility scripts task.
enable_gnr This option is used to enable and configure the use of IBM InfoSphere Global Name Recognition (GNR) with an MDM operational server. This option runs the mpi_gnrconfig.sql file and creates new entries in mpi_libhead, mpi_gnefunc, and mpi_bktxgen database tables. Use this target if you are configuring an algorithm to use the GNRMETA bucket generation function.

disable_gnr This option removes reference to GNR from the mpi_libhead, mpi_gnefunc, and mpi_bktxgen database tables. It does not remove any references to GNRMETA in your algorithm; you must edit your algorithm in InfoSphere MDM Workbench.

You must restart your operational server after you run this option.

generate_perfrpt Takes the results from the Performance Logging Manager process and generates a Performance Log Report. See start_perflogmgr and stop_perflogmgr options.
generate_rocinp Generates the usamp (unmatched samples), dsamp (distribution sampling), and msamp (matched samples) files used by the Threshold Calculator for its calculations. This process can be run from the Generate Threshold Calculator Input Files job in InfoSphere MDM Workbench.

Performance and optional parameters that are found in InfoSphere MDM Workbench can be modified from the command line by using the following Ant properties that are listed here. The equivalent InfoSphere MDM Workbench property follows in parentheses ().

Performance tuning parameters:

mad.rocinputs.threads (Number of threads) (Number of comparison bucket partitions) (Number of random pairs bucket partitions) (Maximum number of input and output partitions)

mad.rocinputs.upair.count (Number of random pairs to generate) (Interval for reporting processed records)

mad.rocinputs.max.bucket.size (Maximum bucket set size)

mad.rocinputs.min.weight (Minimum weight for writing item records) (Number of member partitions)

Optional parameters:

mad.rocinputs.cmpmode (Comparison mode)

mad.rocinputs.use.all.attrs (Skip last step because of too few attributes - true or false)

generate_weights Runs the weight generation process (this process can be run through InfoSphere MDM Workbench).

If you do not want to smooth weights during the generation process, make sure that you set the mad.wgtgen.smooth property to false. For example, -Dmad.wgtgen.smooth=false. The default setting for this property is true.

If mad.wgtgen.smooth is set to true, then a \smoothed subdirectory is created. The directory contains smoothed usamp, msamp, and final binary and text report files. This structure maintains the original, unsmoothed, files that you can compare against the smoother files.

run_jobset Runs a job set that was initially created in InfoSphere MDM Workbench. If InfoSphere MDM Workbench is running on a different computer than the operational server, you must copy the job set XML file from the InfoSphere MDM Workbench project to the operational server. The XML file is at workbench_workspace\project_name\jobTemplates. If the job set includes the Deploy Hub Configuration job, the contents of the entire project must be copied to the server on which the operational server is running.
run_mpitxm Runs the mpitxm utility. For process details, see the mpitxm utility topic.
run_rellinker Runs the relationship linker utility, which creates relationship linkages in bulk fashion.
validate_weights Runs a validation process against the weights generated.