CERN uses Oracle Data Guard standby databases to replace existing servers and storage with new, more powerful models, enabling complete technology refresh with minimal downtime. CERN simply creates a physical standby database on the new technology platform, then executes a Data Guard switchover to move production to the new platform. Downtime is minimized. The new platform is easily tested prior to assuming the production role. If something unforeseen happens after migration, production can be quickly switched back to the original systems without data loss.
The procedure included in the case study covers the simplest possible case using physical standby database, it can be easily modified and extended to minimize downtime during other operations, such as OS upgrades, cluster resize, migration from non-clustered to clustered systems, migration to ASM, and Data Center moves.
For complete, up-to-date information on different platform combinations allowed within a Data Guard configuration please see MetaLink Note 413484.1.
Authors: This case study was written by Maria Girone and Jacek Wojcieszuk of CERN, Dawid Wojcik of CERN Openlab, and Graeme Kerr of Oracle (for Openlab). All observations and recommendations made in this case study are those of the authors.More information about CERN can be found at the CERN website.