Chapter 20. Backup and Recovery

Backing up a system involves copying data from the system’s hard disks onto removable media that can be safeguarded in a secure area. Backing up system data is one of the most crucial system administration functions and should be performed regularly. Backups are used to restore data if files become corrupted or if a system failure or building disaster destroys data. Having a fault-tolerant disk array is not enough. Disk mirroring and RAID 5 protect your data in case of a hardware failure, but they do not protect against file corruption, natural disaster, or accidental deletion of a file. In other words, disk mirroring does not protect against flood damage or fire. In addition, if a program corrupts a particular file, it will be just as corrupt on the mirrored copy as well. Therefore, some type of offsite backup of your data must be in place. Backing up system data—the most important task you will perform—must be done on a regular basis. Although even a comprehensive backup scheme can’t guarantee to prevent loss of information, you can make sure the loss will be minimal.

This chapter describes the methods available to perform a backup, types of backups, development of a solid backup strategy, and restoration of data if you encounter a loss. First you’ll find an explanation of the tar, dd, cpio, and pax commands, which are used to copy data from disk to disk or from disk to tape. Then you’ll learn that the ufsdump and ufsrestore utilities are the preferred method of backing up data from a Solaris system to tape on a regular basis. I’ll also introduce a new method of backing up live file systems called fssnap. Finally, I’ll describe how to back up your entire Solaris operating environment using Web Start archive.

Although many additional third-party backup applications are available that can be purchased to run on Solaris, this chapter describes only those methods available in the standard Solaris 9 distribution.