August 15, 2007, 2:06 PM — Send your Storage question to David Hill today! | See other Storage tips
What seems to be the problem? Some enterprises may spend what seems like a fortune for their disaster recovery architecture, but too often disaster recovery testing is done infrequently at best. Moreover, DR testing is likely to be disruptive, expensive, and inadequate. So IT administrators cannot feel comfortable about their disaster recovery protection infrastructure.
What do you need to know? The software tool cavalry seems to be riding to the rescue for disaster recovery testing. CA XOsoft Assured Recovery has been around for two years. More recently Continuity Software with RecoverGuard and Illuminator with its Virtual Recovery Engine have entered the disaster recovery management space. (Continuity Software recently briefed me as an industry analyst and Illuminator briefed me earlier in the year.) All of these products help with proactive management for disaster recovery management.
The basic problem is that the replicated disaster recovery data center is not necessarily in synchronization with the production data center. Changes to the production data center are not reflected at the replica data center. Dealing with this problem is the task of disaster recovery management products.
Assured Recovery enables a full recovery test of an enterprise's data on the data protection copy. All operations that are necessary to verify the integrity of the data -- and that includes database updates -- are performed. All this is done without impacting the production environment and without having to perform re-synchronization upon the completion of the test.
RecoverGuard and Virtual Recovery Engine take a different tack. (Continuity Software's RecoverGuard will be illustrated for simplicity, but Illuminator is similar.)
RecoverGuard collects information from servers, storage, and databases. For example, servers tell what kind of storage they connect to and databases tell where they keep their data and what they do for disaster recovery. After the information is collected, it is analyzed. Analysis is performed by business service (i.e., application), such as backup, billing, and customer relationship management. Analysis looks for signatures of problems, which can be called gaps (or exposures). The problem may be something such as lack of data consistency, data corruption, or mis-configuration. For example, a snapshot is supposed to protect data on a particular volume on a regular basis, but, for some reason, it is not doing so. The snapshot may have been mis-configured.
Once a problem is detected, an administrator can determine and take corrective action outside the tool. The next cycle (a scheduled time for the information gathering on current status that is followed by analysis) will verify, and then show that the problem has been resolved.
Interestingly enough, even though the software is designed for disaster recovery management, it really helps improve overall IT processes and infrastructure as well as utilizing signatures that are based on vendors' (like Oracle and EMC) best practices. For example, due to configuration problems, some disk volumes are not being used at all although they are supposed to be used! Another example is that one customer was able to solve a production performance issue. Moreover, RecoverGuard can also point out areas of over protection where protection levels exceed the service level agreements (and thereby incur unnecessary costs for excessive resources). Disaster recovery management tools therefore may provide some desirable side-effect benefits.
What can you do about it? Disaster recovery management no longer needs to be considered either impossible at worst or disruptive, inadequate, and overly expensive at best. You now have some tool choices available to help, and the tool that you select may very well provide some desirable side benefits as well.
