A Broken Record

There are some common threads that appear throughout this chapter. Here they are once again:

Mirror the redologs

If all members (or the only member) of an ACTIVE or CURRENT log group are lost, there will be data loss. If the redologs are not mirrored, the chances of this happening are much greater than if the redologs are mirrored. If the redologs are mirrored, the chance that all members of a log group would be damaged is incredibly small. A little research and a little effort up-front will significantly reduce the chances of data loss.

Watch the alert log

Even if the redologs are mirrored, one or more members of a log group may be damaged while the database is operating. The only notification will be an entry in the alert log. Automate the checking of the alert logs for error messages. Otherwise, there may be only one functioning member of a log group, and you may not even know it.

Mirror the control files

A significant portion of this recovery procedure is dedicated to recovering from the loss of all control files. If they are not mirrored, they should be— another example that Prior Proper Planning Purges the Person from the Performance of Painful Procedures.

Use ARCHIVELOG mode

Without ARCHIVELOG mode, Steps 23 and 24 could replace the entire recovery procedure. If one datafile is lost, restore all of them and open the database with the resetlogs option. All changes since the last cold backup will be lost, and a cold backup is the only kind of backup possible, ...

Get Unix Backup and Recovery now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.