Implementing a Restore and Recovery Strategy

Failure / Recovery Scenarios


Implementing a Restore and Recovery Strategy

Phase I - Steps for Diagnosing a Problem

The first phase encompasses researching the nature of the failure. Use V$ dynamic performance views, data dictionary views, trace files, basic operating system commands, and SQL*DBA to determine the problem. 

Phase II - Restore Appropriate Files

Before performing a recovery scenario, it is necessary to determine which file(s) to restore and what state the instance and database must be in to perform the recovery. The objective is to minimise downtime and loss of data, thus recovery of unnecessary data should be avoided.

Phase III - Recover Database

Having determined the nature of the problem and having restored the files if necessary, then the appropriate method of recovery should be performed. Upon completion of the recovery, any proactive measures that can be taken to either prevent or to minimise the same type of failure in the future should be noted.

Phase IV - Back-Up Database

It must be determined if another full offline backup is required. This may be necessary if an incomplete database recovery was performed. A full offline backup, should include the backing up of all data files, control files, to both trace and to a binary copy, and the parameter file.


Failure / Recovery Scenarios

 
Loss of "INACTIVE" Online Redo Log Group

Loss of "Current" Online Redo Log Group

Loss Control Files

Loss of Media

Loss of an Online Rollback Segment Datafile

Loss of System Tablespace Datafile

Loss of Non-System, Non-Rollback Segment Datafile

Corrupted Data Block

Recover from Users Errors

Failure During Hot Backup

Missing Datafile

Loss of a Datafile and Missing Archive Log File

Recover a Lost Datafile with No Backup

Missing Mirrored Online Redo Log Files

Loss of a Control File and Read-Only Tablespace

Loss of Non-Essential Datafile when Database is Down