Trends in Disaster Recovery Practices for Exchange Server

Last week’s discussion of disaster recovery practices for Exchange Server stirred up a lot of interest and feedback from readers. I received quite a bit of email regarding the best practices you're using and the problems you're having.

I saw an overwhelming trend showing the majority of you had already chosen to perform daily full backups. The motivations cited included convenience of a single-tape restore, integrity checking, and simplicity. Single-tape restore and the simplicity this brings are obvious benefits, but the integrity-checking aspect is less obvious. A full backup of the Exchange Information Store (IS) will cause the entire database to go through a per-page checksum process as part of the Exchange backup API. Differential and incremental backups don’t go through the same process because you're backing up only the log files. Daily full backups ensure that you've checked the entire IS for corruption.

Another trend I noticed was that many users are extremely frustrated with their Exchange disaster recovery efforts. Your frustration was directed toward two key areas—backup software vendors and Microsoft. Although I won’t publish specific comments (this is a family newsletter...), it's safe to say that all backup vendors that support Exchange Server need to do a better job of documentation, reliability, and customer support. Many readers shared their sad stories about obtaining support from their backup software vendor after they encounter problems with backing up or restoring Exchange Server. By the way, this is a great litmus test when selecting your backup vendor. Regarding Microsoft, I certainly understand that problems exist. However, I've seen Microsoft put a lot of effort recently into the code, in addition to education, documentation, and support efforts around Exchange Server reliability and disaster recovery. In fact, look for Exchange Server disaster recovery and reliability training at Microsoft Certified Technical Education Centers (CTECs) in the future and specific sessions at MEC 99. Also, look for other vendors to pick up on this trend and provide some great solutions for Exchange.

The last area of concern is that serious misunderstandings still exist about how Exchange disaster recovery really works. I heard from several folks who were trying to perform file-by-file backup and restore operations for their Exchange Servers and were frustrated by the fact that the Exchange files were always open and couldn’t be backed up. If you fall into this category, it's important to understand that online backups are the primary mechanism for Exchange Server, and that there is software (even NT Backup) that supports this function. You should perform offline backups only in special circumstances or as an additional safety measure. Also, if you're using backup software that doesn't have exposure to the Exchange APIs, it's time to upgrade. Most top-tier vendors have support for Exchange online backup APIs and will allow IS backup and recovery without shutting down services.

I hope this information is not news to most readers. If it is, don’t worry—just get up to speed, put some procedures in place, invest in a good solution, and practice, practice, practice.