Topic Last Modified: 2009-02-09
The Microsoft Exchange Disaster Recovery Analyzer tool examines the Exchange database headers to determine the database repair information. Specifically, the Disaster Recovery Analyzer examines the values for the following fields:
- Repair Count
- Old Repair Count
- ECC Fix Success Count
- ECC Fix Error Count
- Bad Checksum Error Count
If the value for a particular field is not zero, the Disaster Recovery Analyzer generates one of the following messages, as appropriate for the particular field:
Database '<DatabaseName>' has been repaired <N> time(s). The last repair time is <RepairDate>.
Old Repair Count
Database '<DatabaseName>' has been repaired <N> time(s) before the last defragmented.
ECC Fix Success Count
An ECC error was fixed and resulted in a good page in database '<DatabaseName>' for <N> time(s). The last log time is <RepairDate>.
ECC Fix Error Count
An ECC error was fixed and resulted in a bad page in database '<DatabaseName>' for <N> time(s). The last log time is <RepairDate>.
Bad Checksum Error Count
A non-correctable ECC/checksum error was found in database '<DatabaseName>' for <N> time(s). The last log time is <RepairDate>.
In Exchange Server 2003 Service Pack 2 (SP2) and in later versions of Exchange, Exchange database headers include additional fields to record repair information. You can use the Eseutil /mh command to display this information. Database repair information may help you determine whether a recurring problem exists with one or more Exchange databases.
The Repair Count value indicates how many database repairs were performed during the last database repair operation.
|The Repair Count field does not indicate how many times the database was repaired. The Repair Count value increases only if an item in the database was repaired during the repair operation. For example, if the Repair Count value is 0, and then two database items are repaired during a single database repair operation, the Repair Count value is set to 2.|
The Old Repair Count field records the Repair Count value since the last offline database defragmentation was performed. A Repair Count value that is greater than the Old Repair Count value indicates that the database has been repaired since the last offline defragmentation operation.
Exchange Server 2003 Service Pack 1 (SP1) and later versions of Exchange include the ability to fix certain error correcting code (ECC) issues during typical database operations. The ECC Fix Success Count field shows how many times the ECC fix operation was successfully performed. Also, when an ECC correction is made to a database page, an additional, more comprehensive test is performed against that page. If this subsequent test fails, the ECC Fix Error Count is incremented by one. Additionally, a -1018 event is logged in the Application log.
The Bad Checksum Error Count field displays the number of -1018 errors in the database. These represent database page errors that could not be repaired automatically.
For More Information
For more information about how to troubleshoot Exchange databases, see TechNet Webcast: Exchange Server Database Troubleshooting and Recovery with the Exchange Troubleshooting Assistant 1.1 (Level 300).
For more information about the ECC-related database headers, see the Microsoft Exchange Team Blog article, Exchange 2003 SP2 Database Header changes.
For more information about Exchange 2007 database recovery, see Disaster Recovery.
For more information about how to use the Eseutil tool, see:
- The Microsoft Exchange Team Blog article, Repairing Exchange databases with ESEUTIL - when and
- The Exchange Help Topic, Eseutil