The process of restoring an RDBMS varies according to the backup methods that you used, of course. How you proceed to restore is based on the status of your nondata and data disks and whether you are able to do partial restores online.
A “data” disk is defined as any disk that contains a database object. (A database object and other terms are defined earlier in this chapter.) If you keep your database objects intact, you actually don’t need to restore the database, you’ve just got to restore all of the parts that make it work! This can range from a restore of a single database setup file to a complete restore of everything on another system.
Your database can’t work if its executables aren’t there. This part of the recovery is much simpler if you have all your executables located in a special filesystem.
In a pinch, you can restore the executables by copying them from a known good system, but that won’t work for your database setup files. Each instance often has an initialization file that sets up certain variables like the instance name and the location of the master database. These setup files usually can be recreated if you have good logs that tell you how you made them the first time, but it is probably easier to recover them from backup.
Databases often require you to edit
system configuration files such as
/etc/system. Changes to these files might include things like customizing shared ...