Enterprises running Exchange 2003 environments must ensure that their existing data can be recovered as smoothly as possible after an interruption. These five Exchange best-practices will provide enhanced recoverability. Follow this advice as appropriate for the enterprise's specific situation.

Recommendations

1. Install multiple Active Directory domain controllers. Because Exchange 2003 stores its configuration data in Active Directory (AD), IT must ensure that data does not become corrupt. A multi-master domain controller architecture, where AD settings replicate to the other domain controllers will mitigate an AD outage. For information on how create the first domain controller for a new child domain, refer to "Step-by-Step Guide to Setting Up Additional Domain Controllers," from Microsoft TechNet.
2. Run daily "full" Exchange backups. Of the Microsoft provided tools, only the full "Exchange aware" file-level backup performs database validation. Industry best practice is to perform full backups at least daily, unless this is obviated by a third-party solution that performs database validation.
3. Backup system configuration. Exchange backups don't capture the system state settings which will be critical for restoring a server from scratch. NTBACKUP can be used to back up system settings, but cannot perform this operation during an Exchange full backup. For more information, refer to "Back up System State data," from Microsoft TechNet.
4. Store transaction log files on separate disk and mirror. Transaction logs are accessed sequentially and can slow down database access. As a tuning tip, placing Exchange logs on a separate drive is a good idea, but as a recovery tip it's an even better idea. Removing Exchange logs from the database drive will help preserve these files which are essential to rebuilding the database after an outage. Mirroring this log partition to a backup location will provide further protection.
5. Run ESEUTIL checks on .EDB files. ESEUTIL allows administrators to determine whether Exchange database files are bad. When a problem arises with the Exchange store, use these key switches for non-destructive testing before doing anything drastic:
   • /g – checks for physical problems in the database. For more information, refer to "How to Run Eseutil /G in Integrity Mode," from Microsoft TechNet.
   • /mh – provides diagnostics and reports the last time/date of database backup. For full details refer to, "Eseutil /M File Dump Mode," from Microsoft TechNet.
   • /k – uses checksums to verify the integrity of each database page (4k segments of the Exchange database). For more information, refer to the Microsoft Knowledge Base article KB825088, "How To: Use the Eseutil Utility to Detect File Header Damage in Exchange 2003."

Bottom Line

Enterprises that recently adopted Exchange 2003 or that plan to delay e-mail server transition until Exchange 2007 SP1 launches will benefit from reviewing their Exchange infrastructures for solid backup and recovery processes. Follow these five tips to strengthen Exchange 2003 recoverability.