Choosing a backup type may seem complicated, but it's not. The bottom line is that the amount of time required for a restore is roughly double the amount of time required to make the backup in the first place. Factor in your RTO to quickly determine how much time you can afford to do a restore, which in turn tells you how long your backup can take if you're going to hit your SLAs and RTO. You can always tweak your backup solution's hardware (for example, by adding more tape drives and striping data across them, or switching to a higher-capacity, faster solution), but the time required for the backup window will ultimately be the number one factor in determining your backup pattern.
You must also include a fudge factor to cover you in case something unexpected happens. Suppose that you actually have only six hours worth of restore window to work with. (In fact, during most restores, IT staffers waste time trying various procedures before they decide that a restore is necessary -- be sure to factor this time into your planning!) Thus, your backup time should be at or below three hours. What kind of backups should you use?
- Full backups take the longest -- assuming that your backup solution can handle 10 to 20 GB per hour, you can restore one server's worth of data in one and a half to three hours -- assuming that nothing goes wrong.
- Incremental backups are smaller, so they take less time to capture and restore. However, they trade time for space; in addition, if the same database page changes more than once over the time span of an incremental set, you'll end up having to play back transactions to change that page over and over, adding to your restore time.
- Differential backups give you easier management and less overhead than incrementals, at the cost of storage growth over time. It's easy to grab a full backup, plus the differential for a given RPO, but you must factor in the time it takes to restore two backups instead of just one.
Without lab testing, it's difficult to pinpoint which combination of full, incremental, and differential backups will best allow you to meet your eight-hour RTO. However, if you have sufficient hardware to support it, daily full backups offer relatively easy restoration, little additional media management overhead, and integrity checking.
Most organizations use a combination of weekly or intra-week daily backups with daily differentials, although your combination may vary. As disk space continues to drop in purchase cost, an increasing number of organizations are doing full backups to disk and intra-day differentials -- doing so gives great coverage at the expense of storage space.
10 tips in 10 minutes: Fundamentals of Exchange Server disaster recovery
Tip 1: Defining Exchange disaster recovery
Tip 2: How Exchange backs up data
Tip 3: Choosing a backup type for Exchange
Tip 4: Online vs. offline Exchange Server backups
Tip 5: Basic Exchange backup and restore
Tip 6: Exchange vendor snapshots and point-in-time copies
Tip 7: VSS for Exchange
Tip 8: Exchange Server replication
Tip 9: Exchange design choices and issues
Tip 10: Exchange disaster recovery planning
This chapter excerpt from the free e-book The Definitive Guide to Exchange Disaster Recovery and Availability, by Paul Robichaux, is printed with permission from Realtimepublishers, Copyright 2005. Click here for the chapter download or download all available chapters here.