Database Backup and

Recovery

Database Management and Administration

Dr Tatiana Simmonds
 Previously . . .
 Described Flashback Technology
 Recovered dropped tables with the Flashback Drop
 Used Flashback Table to restore the contents of a
table to a specific point in the past
 Discussed Flashback Database option
 Introduced Flashback Query

T.Simmonds 2
 Objectives
Discuss Backup and Recovery
Describe different types of failure in the
database
Preparing a database for recoverability
Discuss details of instance recovery
Develop backup strategy
Logical and physical backups
Perform backups
The database recovery process from a backup

T.Simmonds 3
Database Backup and
Recovery

Part1

Database Management and Administration

Dr Tatiana Simmonds
 Problems
 Computer systems fail.
 Hardware breaks.
 Programs have bugs.
 Human procedures contain errors, and people make
mistakes.
 All of these failures can and do occur in database
applications.
 Because a database is shared by many people and
because it often is a key element of an organization’s
operations, it is important to recover it as soon as
possible.
T.Simmonds 5
 What is Backup and Recovery?
In general, backup and recovery refers to the various
strategies and procedures involved in protecting your
database against data loss and reconstructing the
database after any kind of data loss.
Oracle Database provides a sophisticated set of
facilities and utilities for backup and recovery
processing.
 They can be used in many different ways to provide
appropriate backup and recovery for databases,
 ranging from a small workgroup database that can be backed
up when it is unused at night
 to large inter-organisational databases that must be
operational 24 hours per day, 7 days per week (24/7) and can
never be shut down.
T.Simmonds 6
 What is Availability?
Availability is the degree to which an application,
service, or function is accessible on demand
 It is measured by the perception of an application's end
user, but increasingly, availability is measured in dollars,
euro, and yen, not just in time and convenience
Primary characteristics of a highly available solution:
 Reliability
 Recoverability
 Timely error detection
 Continuous operations

T.Simmonds 7
 Backup and Recovery Issues
Oracle Database maintains three types of files that
are important for backup and recovery:
 datafiles
 contain user and system data
 redo log files
 contain logs of database changes
 control files
 small files that store the SCN and the name, contents, and
locations of various files used by a database.
 control files are frequently updated by a database, and they
must be available for a database to be operational

T.Simmonds 8
 What can go wrong?
Database recovery techniques depend on the
type of failure.
Following types of failures can occur in the
database:
 Statement failure
 User process failure
 Network failure
 User error
 Media failure
 Instance failure
T.Simmonds 9
Database Backup and
Recovery

Part2

Database Management and Administration

Dr Tatiana Simmonds
 Statement failure
An individual SQL statement can fail for a number
of reasons:
 Constraint violation
 Insufficient privileges
 Logic errors in applications
 Space management problem

Possible solutions:
 Work with users to validate and correct data
 Provide appropriate object or system privileges
 Work with developers to correct program errors
T.Simmonds 11
 User process failure
A user process may fail for any number of reasons:
 User exits abnormally
 A user’s session is abnormally terminated
 The terminal rebooting
 The program causing an error that
terminates the session
Possible solutions:
 This type of problem is beyond the DBA’s control
 PMON will roll back the transaction, terminate the server
process and release the PGA to the operating system
 Watch for multiple occurrences which might indicate other
problems

T.Simmonds 12
 Network failure
Potential network problems:
 Listener fails
 Network Interface Card (NIC) fails
 Network connection fails

Possible solutions:
 Configure a backup listener and connect-time
failover
 Configure multiple network cards
 Configure a backup network connection
 Work together with the network administrators
T.Simmonds 13
 User error
Users errors used to be the worst errors to manage
User errors are not errors as far as database is
concerned:
 User commits unwanted changes
 User drops the table
Possible solutions:
 To prevent them from occurring in the first place
 software design
 user training
 correct privileges

 Flashback technology
T.Simmonds 14
 Media failure
 A media failure occurs when Oracle Database is unable to
write to a physical file.
 This may occur because of a disk head crash or other disk
failure, because needed devices are not powered on, or because
a file is corrupt.
 Media failure means
 damage to disks, and therefore the files on them
 user error when system or database administrator accidently deleted files
 loss any of on line redo log files, data files or control file

 Possible solutions:
 Prevention and protection
 protect control file and online redo log files through multiplexing
 protect archive log files
 good backup strategy

 Restore and recover the missing files

T.Simmonds 15
 Instance failure
An instance failure is a disorderly shutdown
of the instance, also known as crash
This could be caused by
 a power outage
 switching off or rebooting the server machine
 any hardware problems
 failure of one of the background processes
 shutdown abort command
Possible solutions:
 Instance recovery
T.Simmonds 16
 Instance Recovery
An instance recovery
 is automatic and unavoidable
 is caused by attempts to open a database
whose files are not synchronized on shutdown
 uses information stored in redo log groups to
synchronize files
 involves two distinct operations:
 Rolling forward: Data files are restored to their state before
the instance failed
 Rolling back: Changes that are made but not committed are
returned to their original state
T.Simmonds 17
 Instance Recovery: Details
 On starting an instance SMON process reads the control
file and puts the database into a mount state
 Then SMON checks the file headers of all the datafiles and online
redo log file
 If the file headers are out of sync, SMON starts instance
recovery routine:
 Rolls forward – using the contents of the online redo log files
rebuilds the database buffer cache to the state it was before the
crash
 Opens database
 Rolls back – all uncommitted transactions are rolled back (using
undo)

T.Simmonds 18
Database Backup and
Recovery

Part3

Database Management and Administration

Dr Tatiana Simmonds
 Preparing a Database for Recoverability
 To configure your database for maximum
recoverability:
 The control file must be multiplexed
 The online redo logs must be multiplexed
 The database must be running in ARCHIVELOG
mode, with the archive log files also multiplexed
 There must be regular backups

T.Simmonds 20
 Backup Overview
Backup and recovery is one of the most
important aspects of a database management
A backup strategy provides
a safeguard against data loss
A backup is a copy of data from
your database that can be used to
reconstruct that data.
Backups can be divided into logical and physical
backups, both of which have a number of options
available
T.Simmonds 21
 Logical Backups
 Logical backups of a database contains logical data
such as tables and stored procedures.
 Logical backups exported from a database with an
Oracle export utility and stored in a binary file, for later
re-importing into a database using the corresponding
import utility.
 The records are read independently of their physical location.
 Export file will contain the commands necessary to
completely re-create all the chosen objects and data.
 Data Pump Import utility is used to import the data.

T.Simmonds 22
 Physical Backups
Physical backups are the foundation of any sound
backup and recovery strategy
A physical backup involves copying the files that
constitute the database (datafiles, control files, and
archived redo logs)
Logical backups are a useful supplement to
physical backups but are not sufficient protection
against data loss without physical backups
An effective backup strategy must be based on
physical backups
T.Simmonds 23
 Physical Backup Options
 Backup strategies:
 Whole: A backup of the control file and all datafiles that belong to a
database
 Partial: A portion of the database

 Backup types:
 Full: A non-incremental RMAN backup
 Incremental: Only information that has changed since some previous
backup
 Backups modes:
 Offline: Consistent, also referred to as cold backup
 Online: Inconsistent, also referred to as hot backup

 Backups storage:
 Image copies: Duplicate data and log files in OS format
 Backup sets: Binary, compressed files in Oracle proprietary format

T.Simmonds 24
 Backup Tools
Backups can be performed by using:
Recovery Manager (RMAN)
 is the recommended method of backing up your Oracle
database
Oracle Secure Backup
 complements RMAN existing functionality by adding
backup to tape and network backup capabilities
User-managed backup
 is based on scripts that a DBA must write
 Carried out with OS commands
 is being phased out

T.Simmonds 25
 The Database Recovery Process
Reconstructing the contents of all or part of a
database from a backup typically involves two
phases:
 retrieving a copy of the datafile from a backup,
 reapplying changes to the file since the backup
from the archived and online redo logs, to bring
the database to a desired SCN since the backup
(usually, the present).

T.Simmonds 26
 Essentials
 Discuss Backup and Recovery
 Describe different types of failure in the
database
 Preparing a database for recoverability
 Discuss details of instance recovery
 Develop backup strategy
 Logical and physical backups
 Perform backups
 The database recovery process from a backup

T.Simmonds 27