Using USMT in Microsoft Deployment Toolkit

User state migrations can be started and controlled in a number of ways. Among these are direct command-line execution, scripting, MDT 2010, and Configuration Manager. The section titled "Understanding USMT Components" earlier in this tutorial describes the command-line options for running USMT directly or driving it by using scripts. This section describes how to enable USMT in MDT 2010, as well as how to add custom migration .xml files to MDT 2010.

State Migration in MDT 2010

In previous tutorial "Developing Disk Images," describes the task sequence and Task Sequencer that MDT 2010 uses for deploying Windows 7. The default task sequence separates the process into two phases. One of the preinstallation phases is State Capture; one of the postinstallation phases is State Restore. The entire state migration work is tucked into these two phases.

In the State Capture phase, the Capture User State step runs ZTIUserState.wsf /Capture to capture user state. It uses settings from the deployment share's CustomSettings.ini file or the MDT 2010 database. In the State Restore phase, the Restore User State step runs ZTIUserState.wsf /Restore to restore the state data captured in the Capture User Step.

For the /capture command-line option, ZTIUserState.wsf reads its settings (UDDShare, UDDir, and so on) from the environment and then chooses the best place to create the data store based upon UserDataLocation. In the final step, the script executes ScanState with the command-line arguments that it assembled from the data in the environment, adding the command-line options for hard-link migration. For the /restore command-line option, ZTIUserState.wsf retrieves information about the data store it created from the environment and then runs LoadState using the command line that it assembled from that information, also adding commandline options for hard-link migration.