Acrobat Connect and database mirroring with Microsoft SQL Server 2005
August 20th, 2007
One of the new features in SQL Server 2005 is database mirroring — the ability to replicate data in real time from a primary database to a mirror, or backup, database. There were ways to accomplish this in SQL Server 2000, but they had significant performance problems. As of SQL Server 2005 SP1, mirroring is production-ready and performs well.
There’s just one slight problem with using this feature with Acrobat Connect — the JDBC driver that Connect uses is not mirroring aware. So you can mirror your database, but you can’t take advantage of the automated failover features built into SQL Server 2005, because Connect’s JDBC driver can only see the primary database. In the event of a primary database failure, you need to bring the mirror online, then go into the Connect console manager and change the database settings to now point to the mirror, and finally restart Connect. This is a time-consuming process. For example, there is one large Connect installation, with a dedicated IT team managing Connect, where this process can take half an hour or more, during which time Connect is entirely offline.
At some point, the Connect JDBC drivers will be updated to a newer version that is mirroring-aware. Until then, Connect administrators will be able to use mirroring only for “warm backup.” This means having a system administrator and a DBA on call and available 24/7 to perform the failover.
Or, of course, you can do what we recommend, which is to use ConnectSolutions to manage it all for you 
By optimizing the failover process, we are able to move a Connect installation from the primary database to the mirror database so fast that meeting users will notice only a short “blip”, before their meeting continues as before. And we do this 24/7, 365 days of the year. It’s all part of our fanatical focus on being the very best, bar none, at managing Connect.
Leave a Reply