Databases Reference
In-Depth Information
Recovery Manager Restore
and Recovery Techniques
Once a backup has been taken the second part of the process is to perform the restore
or recover operation. This is an operation no one wants to have to undertake, but you
must be prepared to do so if it is required. This chapter deals with the recovery tasks
from the recovery manager point of view.
All the scenarios assume that you have a database in archivelog mode and a valid
backup which was taken using recovery manager. The database used to perform the
demonstrations is a regular general purpose database created with the DBCA.
The scenarios are performed with the SYSDBA role because this role can manage
the Oracle instance and it can perform incomplete recovery operations.
Oracle database recovery
There are two situations when a database recovery process will be required, one is
after an instance failure, and the other is after a media failure. The recovery process
takes information from the redo log files or the archivelog files as required and applies
the changes found there against the datafiles. The applied changes depend on each
individual datafile, and they range from the last applied change against the datafile to
the last
System
Change
Number
(
SCN
) recorded at the control file.
The recovery process will read the transactions and it will apply all the recorded
changes against the datafiles, the changes are from either committed or uncommitted
transactions. During the first recovery phase all changes stored in either the archive
log files or the redo log files are applied against the datafiles (this is known as rolling
forward). During the second recovery phase all uncommitted transactions are rolled
back. The recovery process ends when all the changes have been applied against the
database files or if the DBA explicitly defined a point where the recovery process must
stop, then the recovery process will end when the recovery manager reaches that point.
Search WWH ::
Custom Search