Performing Availability Testing
You should test your system for disaster recovery to verify its ability to recover from various levels of failure, ranging from small-scale failures (such as a network card failure) to the loss of a production server. Your disaster recovery testing should include the following steps:
Test individual hardware component failure.
You should test the ability of the system to recover from an individual hardware component failure, such as a network or disk.
Test single BizTalk server failure.
In most BizTalk Server production environments, host processing is spread out among multiple computers running BizTalk Server in a single BizTalk group. What is the ability of the BizTalk group to continue host processing in the event one of the servers in the group fails?
Test cluster node failover.
If Windows Clustering is used to provide high availability for the BizTalk Server databases or BizTalk Hosts, you should verify cluster node failover functionality. For more information about using Windows Clustering to provide high availability for BizTalk Server, see Checklist: Providing High Availability with Fault Tolerance or Load Balancing.
Test recovery of BizTalk Server databases using log shipping.
You should verify the recovery of the BizTalk Server databases. For more information about using log shipping to back up and restore BizTalk Server databases, see What Is BizTalk Server Log Shipping? in this guide or Log Shipping (http://go.microsoft.com/fwlink/?LinkID=153450) in the BizTalk Server 2009 Help.
For a checklist of steps that you should follow to increase the availability of a BizTalk Server environment using disaster recovery, see Checklist: Increasing Availability with Disaster Recovery.