Disaster Recovery Strategy

If you implement the root and file replication capabilities provided by Dfs, you can protect yourself against the loss of the server hosting a Dfs root or a server that contains the shared folders themselves. However, this does not eliminate the need to back up at least one copy of all your data. Remember that automatic replication can propagate corrupted information as well as valid data. When you repair or replace the server that has failed, you might have to restore the Dfs topology or shared data stored on it.

Therefore, it is important to plan for disaster recovery of the Dfs namespace. As noted earlier in "Replication Strategy," there are two areas of concern with Dfs:

• Saving your data — that is, the shared folders in the Dfs namespace.

• Saving the Dfs topology.

  • For stand-alone Dfs, this means backing up the Dfs configuration stored in the registry.

  • For domain-based Dfs, this means backing up the Dfs configuration stored in Active Directory and in the registry of each server.

For more information about backing up and restoring replicated files, see "File Replication Service" in this book.

For more information about backing up and restoring Active Directory, see "Active Directory Backup and Restore" in this book.

For more information about system recovery, see "Repair, Recovery, and Restore" in the Microsoft ® Windows ®  2000 Server Resource Kit Server Operations Guide .

Note

The dfscmd command contains a BATCH option that can be used to back up and restore the Dfs topology you have configured on a root server to a floppy disk. For more information, see "Scripting" earlier in this chapter.