Release 10.1A: OpenEdge Development:
Progress Dynamics Administration

Tips for customized DCU sessions

Remember the following points while customizing a DCU session for your deployment:

Follow the flow of the standard DCU session. Collect all the information you need to complete the deployment first, then start processing. This approach will become mandatory in future versions of the DCU.
The order in which you make changes is critical. For example, applying datasets can overwrite existing user data. If you must save the existing data, run a program to move the data to a temporary storage area before applying the datasets that would overwrite the data. Then run another program to add the saved data back to the appropriate tables.
Because the full Progress Dynamics environment is not available during Phase 1, it is impossible to execute schema triggers when data is written. Procedures running during this phase should alter as few records and as little data as possible.
In general, the only upgrade procedures that should run during Phase 1 are those needed to migrate data from one schema version to another.
Test how each change to your application affects your deployment. Consider changing the application design to avoid complex fixes during deployment. Designing your application and deployment should be part of the same iterative cycle.
Examine all the patch levels for a given deployment for potential conflicts. Each time you add a new patch procedure or ADO, you must verify that it does not cause problems for any of the earlier patches.
In general, you should use ADOs for your deployments. When possible, run all fix programs on a given set of data in your central development Repository. Then generate the ADOs for your deployment. You should avoid using fix programs in deployments whenever possible.