Last week, I talked about Exchange Server's scheduled maintenance tasks. Aside from these tasks, which occur automatically, which manual tasks should you perform regularly, and which tasks should you perform only in the case of an emergency?
Let's start with backups. We all know we're supposed to perform regular backups, and most of us do so. Exchange Server 2003 likely will change the way we perform backups, though. Microsoft Volume Shadow Copy Service (VSS), with its speed and convenient point-in-time copies that don't tie us to any specific vendor, promises to usurp our old friend the tape drive. (Only time will tell how well VSS will live up to its promise.) And Exchange 2003's Recovery Storage Group feature offers an immediate improvement to the way we perform restores. Recovery Storage Groups are particularly useful for restoring deleted messages or mailboxes: If you can use Recovery Storage Groups in your organization, you might not need to do mailbox-level backups.
Let's move on to running Isinteg and Eseutil. These tools check the logical and physical structure of your database's .edb and .stm files, respectively. Typically, you should run the tools only when you suspect a problem with a database; running the tools regularly wastes a lot of time, especially because you can run them only on an unmounted database. The way to discover database problems is by reviewing your event logs--and you should perform that task on a regular basis.
What about performing an offline defrag of your database? Doing so is necessary only when you're extremely tight on space. Because you must take databases offline during defragmentation, this "preventive" task actually causes downtime. Instead, the next time your servers are down for maintenance, install a FireWire (IEEE 1394) card and buy a couple of large FireWire disks. Presto! No more running out of disk space (which always seems to happen at the worst times).
Another task that ranks right up there with cleaning out your gutters each spring and fall is cleaning out the Badmail folders on your SMTP gateways. The messages in these folders can take up a surprising amount of space, and (unless you're morbidly curious) you've no reason to keep such mail around. While you're rooting around in the file system, you might as well have a look at the message-tracking log-retention settings for your servers and ensure that you aren't keeping log files longer than necessary. If you run Exchange 2000 Server, also check the permissions on the tracking.log shares; the default permissions are rather loose and might allow unwanted browsing of the tracking logs.
No discussion of regular Exchange maintenance would be complete without mentioning monitoring. Most Exchange administrators use some kind of monitoring infrastructure, even if it's only the Exchange monitoring tools. Be sure that your system is configured to send you alerts when necessary.
These kinds of basic maintenance tasks aren't as glamorous as, say, migrating a 20,000-seat organization from stone tablets to Exchange 2003. But as I learned a long time ago in the Marines, if you take care of your gear, it'll take care of you.
