Chapter 9. Dynamic cube lifecycle 253
When the reports achieve a satisfactory level of performance, the cube, package, and reports
can be deployed into the production environment.
The production environment is likely to be a highly controlled environment. The changes that
occur in this environment should be well tested before they are implemented at this level.
The final set of in-database aggregate tables must be created in the production environment.
If the dynamic cube is transferred by using a deployment archive from the test environment, it
must already contain the in-memory aggregates. If the dynamic cube is transferred by using
the Cognos Cube Designer, then an administrator will need to use DQA to apply the
in-memory aggregates to the production environment. If security is a requirement for the
system, the administrator will also map users and groups to the security views.
Even if the test and production environment have mirror databases, a good idea is to make
sure that all reports still run in the expected amount of time before going live.
9.3 Best practices for deploying between environments
This section contains various hints and tips to facilitate movement between environments
during the dynamic cube lifecycle.
9.3.1 Keep a copy of the dynamic cube project file under source control
Consider the dynamic cube project file as the original version of the cube model. When the
Cognos Cube Designer publishes a dynamic cube as a data source, it transforms the cube
model in the dynamic cube project into a metadata format that the Dynamic Cube Server will
use at run time. However, this transformation is one-way. There are no means to
reverse-engineer a dynamic cube data source back into a dynamic cube project.
While the transformed metadata resides within the content store and can be transferred from
machine to machine through a deployment archive, that cube metadata remains static. To
make any changes to the design of the cube model, the modeler needs to update the original
dynamic cube project from within the Cognos Cube Designer and publish it again. If the
project file is ever lost, the modeler will need to re-create the cube from the beginning to
modify the cube. Therefore, it is important to keep track of the project file even after a cube
goes into production.
Keeping the project file under source control ensures that the file is backed up and helps to
manage changes to the model, when multiple modelers are working on the same cube.