Automated Database Refresh in Very Large and Highly Replicated Environments
E. M. King · ePublications at Regis University (Regis University) · 2011
Refreshing non-production database environments is a fundamental activity. A non-productive environment must closely and approximately be related to the productive system and be populated with accurate, consistent data so that the changes before moving into the production system can be tested more effectively. Also if the development system has more related scenario as that of a live system then programming in-capabilities can be minimized. These scenarios add more pressure to get the system refreshed from the production system frequently. Also many organizations need a proven and performant solution to creating or moving data into their nonproduction environments that will neither break business rules, nor expose confidential information to non-privileged contractors and employees. But the academic literature on refreshing non-production environments is weak, restricted largely to instruction steps. To correct this situation, this study examines ways to refresh the development, QA, test or stage environments with production quality data, while maintaining the original structures. Using this method, developer's and tester's releases being promoted to the production environment are not impacted after a refresh. The study includes the design, development and testing of a system which semi-automatically backs up (saves a copy of) the current database structures, then takes a clone of the production database from the reporting or Oracle Recovery Manager servers, and then reapplies the structures and obfuscates confidential data. The study used an Oracle Real Application Cluster (RAC) environment for refresh. The findings identify methodologies to perform the refresh of non-production environments in a timely manager without exposing confidential data, and without over-writing the current structures in the database being refreshed. They also identify areas of significant savings in terms of time and money that can be made by keeping the structures for developers with freshened data.