XADM: How Log Files Are Replayed During Exchange 2000 Server Restore Procedure (232922)

SUMMARY

Microsoft Exchange 2000 Server supports multiple databases on each server. Information pertaining to Exchange Server databases that was stored in the registry in earlier versions of Exchange Server is now located in the Windows 2000 Active Directory (that is, database paths).

Exchange 2000 also adds support for multiple storage groups, as well as multiple instances of Extensible Storage Engine (ESE) on each server. Each instance of ESE manages log files for a particular storage group (databases within the same storage group share the same set of log files). There is also one reserved instance of ESE specifically for online backup and restore. This allows the backup and restore of a single Exchange 2000 database.

MORE INFORMATION

The following is a description of the log file's replay logic for a single database restore:

The database gets copied to its original location (ESE checks the * Restore.env file for this information) and the log files from tape get copied to the temporary folder specified during the restore operation ("Log Path" box).
The ESE instance used for the restore process replays these log files to the database (this only happens if you click to select the Last Backup Set check box during the restore operation), then checks the ESE instance running against the storage group to determine which production log files it needs to replay next.
Before starting the replay process, ESE checks the log file sequence and warns you if a log is missing (or doesn't match the logs from the backup set). In this case, ESE only replays the logs from tape and brings the database to a consistent state.

NOTE: Data created since the restore will be lost unless you can provide the missing log files.

You need to mount the database manually unless you have clicked to select the Mount Database After Restore check box.