SYMPTOMS
The standard transaction log recovery process may not replay transaction log files that are in the Exchsrvr\Mdbdata log folder, even though the following events indicate successful replay:
Event Type: Information
Event Source: ESE98
Event Category:
Event ID: 301
Date: 10/17/2001
Time: 5:52:11 AM
User: N/A
Computer: server_name
Description:
Information Store (1728) The database engine has begun replaying logfile H:\Apps\Exchange\MDBDATA\E0014553.log.
For more information, click http://search.support.microsoft.com/search/?adv=1.
CAUSE
This recovery process behavior occurs by design. During the recovery procedure, the information store process queries Active Directory for the existence of the database globally unique identifier (GUID) that is represented by the transaction log files in the Mdbdata log folder. Because this is an off-line recovery process, the information store does not have access to the production Active Directory, and therefore cannot locate the GUID for the associated databases. At this point, these logs are determined to be from a different database; therefore, the soft recovery process does not commit these transactions.
If you run this recovery on a production server, or on a recovery server that has access to a copy of the production Active Directory, the soft recovery process can validate the database GUID and the logs are replayed as expected. The following events indicate which logs are replayed during the recovery process.
NOTE: The event ID 204 information message indicates the beginning of log replay.
The event ID 204 information message is similar to the following:
Event Type: Information
Event Source: ESE98
Event Category:
Event ID: 204
Date: 10/17/2001
Time: 5:49:01 AM
User: N/A
Computer: server_name
Description:
Information Store (1728) The database engine is restoring from backup. Restore will begin replaying logfiles in folder H:\Apps\Exchange\TempLog\SG1\ and continue rolling forward logfiles in folder H:\Apps\Exchange\TempLog\SG1\.
The event ID 204 information message is followed by a series of ESE98 event ID 301 information messages (similar to the following) that describe the exact log files that are being recovered:
Event Type: Information
Event Source: ESE98
Event Category:
Event ID: 301
Date: 10/17/2001
Time: 5:49:09 AM
User: N/A
Computer: server_name
Description:
Information Store (1728) The database engine has begun replaying logfile H:\Apps\Exchange\TempLog\SG1\E001453A.log.
When log recovery is complete, an ESE98 event ID 205 information message that is similar to the following is logged:
Event Type: Information
Event Source: ESE98
Event Category:
Event ID: 205
Date: 10/17/2001
Time: 5:52:10 AM
User: N/A
Computer: server_name
Description:
Information Store (1728) The database engine has stopped restoring.
At this point, several more ESE98 event ID 301 information messages may be logged. Then an ESE98 event ID 494 error message that is similar to the following is logged, which indicates that the database GUID that is represented in the log files cannot be located in Active Directory, and that these logs will not be recovered:
Event Type: Error
Event Source: ESE98
Event Category:
Event ID: 494
Date: 10/17/2001
Time: 5:54:56 AM
User: N/A
Computer: server_name
Description:
Information Store (1728) Database recovery failed with error -1216 because it encountered references to a database, 'F:\Apps\Exchange\MDBDATA\MBX-AUSXMBT102VS1-SG1-2.edb', which is no longer present. The database was not brought to a consistent state before it was removed (or possibly moved or renamed). The database engine will not permit recovery to complete for this instance until the missing database is re-instated. If the database is truly no longer available and no longer required, please contact PSS for further instructions regarding the steps required in order to allow recovery to proceed without this database.
Finally, an ESE98 event ID 454 error message that is similar to the following is logged to complete the process:
Event Type: Error
Event Source: ESE98
Event Category:
Event ID: 454
Date: 10/17/2001
Time: 5:54:56 AM
User: N/A
Computer: server_name
Description:
Information Store (1728) Database recovery/restore failed with unexpected error -1216.
WORKAROUND
To work around this behavior and replay the logs, you need to copy all of the transaction logs that you want to recover to the temporary restore folder that you defined in the restoration process. To copy all of the transaction logs that you want to recover to the temporary restore folder and then replay the logs:
- Restore all of the online backups to the recovery server.
IMPORTANT: Do not select the Last Backup Set check box. This allows you to copy all of the transaction logs to the temporary restore folder before you start the manual replay process. - Copy all of the offline log files to the temp_restore_folder\storage_group_name folder, where temp_restore_folder is the temporary restore folder that you defined in the restoration process, and storage_group_name is the name of the storage group.
- Make sure that the log files are contiguous. Log replay only plays forward as long as the sequence is contiguous; when this sequence is interrupted, replay stops.
- After you are satisfied with the state of the log files, change directory to the temp_restore_folder\storage_group_name folder, and then run either of the following commands:
c:\exchsrvr\bin\eseutil /cc
-or-
c:\program files\exchsrvr\bin\eseutil /cc
After this process is complete, you can manually mount the databases that you restored. In the event log, an ESE98 event ID 204 information message is logged to indicate the start of recovery, and then a series of ESE98 event ID 103 information messages are logged. These information messages indicate the exact log numbers that are recovered. Finally, an ESE98 event ID 205 information message is logged to indicate the completion of the recovery process.