When the primary data center fails, the second data center is activated as a replacement of the primary site. When the second data center is activated, the original data center configuration remains in Active Directory. This configuration is used by the second data center after it has been activated.
SCR only applies to providing messaging service redundancy during a site disaster. When the passive node is brought online and the SCR is set up properly, hosted users will continue sending and receiving e-mails as usual. However, service providers need to establish additional measurements to support other services in the Hosted Messaging and Collaboration, such as provisioning, Microsoft Windows SharePoint Services®, and monitoring and reporting.
In the Hosted Messaging and Collaboration solution, when mailboxes are provisioned against a pool of mailbox stores, Resource Manager tracks the usage of the resources and maintains a link in the Resource Manager Database. More specifically, the GUID of the Exchange 2007 information store on which the mailbox was allocated is tracked. In a SCR deployment, the failover information store is separate from the primary information store, and has a different GUID in the Active Directory. When mailboxes fail over to the standby node, that GUID that glues the Resource Manager to the Exchange information store is no longer valid and will cause failures when new mailbox are created or existing mailboxes are modified through provisioning. If service providers plan to implement SCR, they need to design a plan to support provisioning failover to and failback from the remote site, including replicating the provisioning data between the primary site and secondary site. Additionally, the link in the Resource Manager Database must be updated appropriately. There is a feature planned in the Hosted Messaging and Collaboration version 4.5 that will make the update easier.
For Windows SharePoint Services disaster recovery, database mirroring is recommended to provide more reliability within Microsoft Office SharePoint Server. For more information about Database Mirroring with Office SharePoint Server and Windows SharePoint Services, see Using Database Mirroring with Office SharePoint Server and Windows SharePoint Services
The monitoring and reporting disaster recovery should be broken into two segments:
- Microsoft Operations Manager (MOM) / System Center Operations
Manager 2007 operations manager database recovery
- Performance Warehouse Database (PWDB) recovery (reporting).
Since this a Microsoft SQL Server® database, disaster recovery
planning described by the SQL Server team should be sufficient.
Once the database is restored, use the deployment guides to get
reporting going again. For more information, see INF: Disaster Recovery Planning for SQL
Server.
For more information about how to do disaster recovery for non-MBX roles and achieving high availability for data and service, see High Availability.