Related Extract Engines
While File-AID/EX provides two Related Extract engines for SQL Server and Oracle databases, by default these databases use the Execution Server-Driven Engine. The Database Server-Driven Engine, however, uses a much higher degree of parallel processing and generally provides significantly better extract performance.
The Database Server-Driven Engine divides the extraction into smaller "data preparation" and "data processing" tasks. Data preparation tasks involve executing SQL queries on the database server and creating and deleting tables for temporary storage of data. These tasks can require significant CPU times and available physical memory. To regulate usage of the database server resources, adjust the data preparation threads settings. Data processing tasks focus on actual data retrieval, application of data privacy rules, and writing the data to the extract files; they may require significant resources on the execution server, disk drives, and files servers to which extracted files are written. To adjust the load on these resources, change the data processing threads settings.
Considerations
The higher degree of parallel processing of the Database Server-Driven Engine can increase the changes of exceeding the scalability limits of the database server or the execution server and can affect response time for other clients accessing the same database server. If this occurs, reduce the number of threads to a lower value to lower the resource consumption.
While the extracted data is guaranteed to have referential integrity, the parallelization can cause some results inconsistencies if the extract specification contains circular relationships or if the extract scope excludes the sibling relationships or indirect dependent relationships. If this is a concern, set the data preparation threads to 1 or include both sibling and indirect dependent relationships.
In rare cases, tables created by the Related Extract engine for temporary storage of data may not be deleted, for example if the execution server loses connectivity with the database. These tables, which are all prefixed with FACS_RE_, can be easily deleted using the File-AID DB Cleanup Utility or another means.