I was watching some of the online training you can access here on SSWUG. Denny Cherry has a series on Windows, SQL Server and SANS with the goal of helping create an installation that is fault tolerant and performant.
One of my favorite parts of the training is when Denny reminds us that our Disaster recovery plan is much more useful when an actual disaster is emulated, and the plan is validated.
Many times we have a plan based on scenarios where the primary persons responsible for a system are not available during the restoration process. What happens if a disaster happens where not only is your main site disabled for some reason, but the indviduals that maintain the site are also not available. A broad enough power outage for a long enough period of time could disable your data center as well as telephones. Should your backup site be located too far away, secondary people will need to restore your systems and get things running again.
The point is that your recovery exercise should be executed at some time without support from the people that put together the plan and maintain things day to day. If you can restore your systems under those conditions, then you have a great deal of security in the success of recoverying in a real disaster.
How far should you take your tests? Should you actually failover assuring that the necessary resources are available? Are there other things you’d like to add to the conversation. Leave your comments here online or drop an Email to btaylor@sswug.org.
Cheers,
Ben