Skip Headers
Oracle® Database Backup and Recovery User's Guide
12c Release 1 (12.1)

E50658-05
Go to Documentation Home
Home
Go to Book List
Book List
Go to Table of Contents
Contents
Go to Index
Index
Go to Master Index
Master Index
Go to Feedback page
Contact Us

Go to previous page
Previous
Go to next page
Next
PDF · Mobi · ePub

9 Backing Up the Database

This chapter explains how to perform the most basic backup tasks and implement backup strategies using RMAN. This chapter contains the following topics:

See Also:

Overview of RMAN Backups

This section provides an overview of RMAN backups.

Purpose of RMAN Backups

The primary purpose of RMAN backups is to protect your data. If a media failure or disaster occurs, then you can restore your backups and recover lost changes.

You can also make backups to preserve data for long-time archival, as explained in "Making Database Backups for Long-Term Storage", and to transfer data, as explained in the chapters included in Part VII, "Transferring Data with RMAN".

Basic Concepts of RMAN Backups

As explained in Chapter 8, "RMAN Backup Concepts," you can back up all or part of your database with the BACKUP command from within the RMAN client.

In many cases, after your database has been configured in accordance with your backup strategy, you can back up the database by entering the following command at the RMAN prompt:

RMAN> BACKUP DATABASE;

RMAN uses the configured settings, the records of previous backups, and the control file record of the database structure to determine an efficient set of steps for the backup. RMAN then performs these steps.

As explained in "RMAN File Management in a Data Guard Environment", you can run RMAN backups at any database in a Data Guard environment. Any backup of any database in the environment is usable for recovery of any other database if the backup is accessible. You can offload all backups of database files, including control file backups, to a physical standby database and thereby avoid consuming resources on the primary database.

See Also:

Specifying Backup Output Options

If you specify only the minimum required options for an RMAN command such as BACKUP DATABASE, then RMAN automatically determines the destination device, locations for backup output, and a backup tag based on your configured environment and built-in RMAN defaults.

You can also provide arguments to BACKUP to override these defaults. The most typical options are described in the following sections:

See Also:

Chapter 10, "Backing Up the Database: Advanced Topics" to learn about advanced backup options such as duplexing and restarting backups

Specifying the Device Type for an RMAN Backup

The BACKUP command takes a DEVICE TYPE clause that specifies whether to back up to disk or tape device. Example 9-1 illustrates a backup to disk.

Example 9-1 Specifying Device Type DISK

BACKUP 
   DEVICE TYPE DISK
   DATABASE;

When you run BACKUP without a DEVICE TYPE clause, RMAN stores the backup on the configured default device (disk or SBT). You set the default device with the CONFIGURE DEFAULT DEVICE TYPE command described in "Configuring the Default Device for Backups: Disk or SBT".

Specifying Backup Set or Copy for an RMAN Backup to Disk

RMAN can create backups on disk as image copies or as backup sets. "Configuring the Default Type for Backups: Backup Sets or Copies" explains how to configure the default disk device. You can override this default with the AS COPY or AS BACKUPSET clauses. To back up to disk as image copies, use BACKUP AS COPY as shown in Example 9-2.

Example 9-2 Making Image Copies

BACKUP AS COPY
  DEVICE TYPE DISK 
  DATABASE;

To back up your data into backup sets, use the AS BACKUPSET clause. You can allow backup sets to be created on the configured default device, or direct them specifically to disk or tape, as shown in Example 9-3.

Example 9-3 Making Backup Sets

BACKUP AS BACKUPSET 
  DATABASE;

BACKUP AS BACKUPSET 
  DEVICE TYPE DISK 
  DATABASE;

BACKUP AS BACKUPSET 
  DEVICE TYPE SBT 
  DATABASE;

Specifying a Format for RMAN Backups

RMAN provides a range of options to name the files generated by the BACKUP command. RMAN uses the following set of rules to determine the format of the output files, which are listed in order of precedence:

  1. If a FORMAT parameter is specified on the BACKUP command, then this setting controls the generated file name.

    For example, you can direct the output to a specific location, as shown in the following command:

    BACKUP DATABASE 
      FORMAT "/disk1/backup_%U";  # specifies a location on the file system
    

    In this case, backups are stored with generated unique file names with the prefix /disk1/backup_. The %U substitution variable, used to generate a unique string at this point in the file name, is required.

    You can also use the FORMAT parameter to name an ASM disk group as the backup destination, as shown in the following example:

    BACKUP DATABASE 
      FORMAT '+dgroup1';  # specifies an ASM disk group
    

    No %U is required in this case because Automatic Storage Management (ASM) generates unique file names as needed. However, you can specify %U if desired.

    Note:

    If you specify FORMAT when a fast recovery area is enabled, then RMAN obeys the FORMAT setting. If no location is specified in the FORMAT clause, then RMAN creates the backup in a platform-specific location.
  2. If a FORMAT setting is configured for the specific channel used for the backup, then this setting controls the generated file name.

  3. If a FORMAT setting is configured for the device type used for the backup, then this setting controls the generated file name.

  4. If a fast recovery area is enabled during a disk backup, and if FORMAT is not specified, then RMAN creates the backup with an automatically generated name in the fast recovery area.

  5. If no other conditions in this list apply, then the default location and file name format of the backup are platform-specific.

See Also:

Oracle Database Backup and Recovery Reference to learn about the FORMAT clause, and the installation guides in the Oracle Database documentation library to learn about the default file locations for your platform

Specifying Multiple Formats for Disk Backups

Typically, you do not need to specify a format when backing up to tape because the default %U variable generates a unique file name for tape backups. When backing up to disk, however, you can specify a format to spread the backup across several drives for improved performance. In this case, allocate one DISK channel for each disk drive and specify the format string on the ALLOCATE CHANNEL command so that the file names are on different disks. For example, issue the following command:

RUN
{ 
  ALLOCATE CHANNEL disk1 DEVICE TYPE DISK FORMAT '/disk1/%d_backups/%U';
  ALLOCATE CHANNEL disk2 DEVICE TYPE DISK FORMAT '/disk2/%d_backups/%U';
  ALLOCATE CHANNEL disk3 DEVICE TYPE DISK FORMAT '/disk3/%d_backups/%U';
  BACKUP AS COPY DATABASE; 
} 

You can distribute backups in this manner by default in the future, by configuring channels as follows:

CONFIGURE DEVICE TYPE DISK PARALLELISM 3;
CONFIGURE DEFAULT DEVICE TYPE TO DISK;
CONFIGURE CHANNEL 1 DEVICE TYPE DISK FORMAT '/disk1/%d_backups/%U';
CONFIGURE CHANNEL 2 DEVICE TYPE DISK FORMAT '/disk2/%d_backups/%U';
CONFIGURE CHANNEL 3 DEVICE TYPE DISK FORMAT '/disk3/%d_backups/%U';
BACKUP AS COPY DATABASE;

Specifying Tags for an RMAN Backup

RMAN attaches a character string called a tag to every backup it creates, as a way of identifying the backup. You can either accept the default tag or specify your own with the TAG parameter of the BACKUP command.

About Backup Tags

User-specified tags are a useful way to indicate the purpose or usage of different classes of backups or copies. You can tag backup sets, proxy copies, data file copies, or control file copies. For example, you can tag data file copies that you intend to use in a SWITCH command as for_switch_only and file copies to use only for a RESTORE command as for_restore_only.

Tags do not need to be unique, so multiple backup sets or image copies can have the same tag (for example, weekly_backup). Assume that you specify that a data file be restored from backups that have a specific tag. If multiple backups of the requested file have the desired tag, then RMAN restores the most recent backup that has the desired tag, within any constraints on the RESTORE command.

In practice, tags are often used to distinguish a series of backups created as part of a single strategy, such as an incremental backup strategy. For example, you might create a weekly incremental backup with a tag like BACKUP TAG weekly_incremental. Many forms of the BACKUP command let you associate a tag with a backup, and many RESTORE and RECOVER commands let you specify a tag to restrict which backups to use in the RESTORE or RECOVER operation.

If you do not explicitly specify a tag with the TAG parameter of the BACKUP command, then RMAN implicitly creates a default tag for backups (except for control file autobackups). The format of the tag is TAGYYYYMMDDTHHMMSS, where YYYY is the year, MM is the month, DD is the day, HH is the hour (in 24-hour format), MM is the minutes, and SS is the seconds. For example, a backup of data file 1 may get the tag TAG20070208T133437. The date and time refer to when RMAN started the backup in the time zone of the instance performing the backup. If multiple backup sets are created by one BACKUP command, then each backup piece has the same default tag.

Tags are stored in uppercase, regardless of the case used when entering them. The maximum length of a backup tag is 30 bytes. Tags cannot use operating system environment variables or use special formats such as %T or %D.

See Also:

Oracle Database Backup and Recovery Reference for the default format description in BACKUP ... TAG

Specifying Tags for Backup Sets and Image Copies

The characters in a tag must be limited to the characters that are legal in file names on the target database file system. For example, Automatic Storage Management (ASM) does not support the use of the hyphen (-) in the file names it uses internally, so a tag including a hyphen (such as weekly-incr) is not a legal tag name for backups in ASM disk groups.

When you tag a backup set, the tag is an attribute of each backup piece in a given copy of a backup set. If you create a multiplexed backup set, then each copy of the backup set is assigned the same tag. Example 9-4 creates one backup set with the tag MONDAYBKP.

Example 9-4 Applying a Tag to a Backup Set

BACKUP AS BACKUPSET
  COPIES 1 
  DATAFILE 7
  TAG mondaybkp; 

When you specify a tag for image copies, the tag applies to each individual copy. Example 9-5 shows that copies of data files in tablespaces users and tools are assigned the tag MONDAYCPY.

Example 9-5 Applying Tags to Image Copies


BACKUP AS COPY 
  TABLESPACE users, tools
  TAG mondaycpy;

You can use FROM TAG to copy an image copy with a specific tag, and then use TAG to assign the output copy a different tag. Example 9-6 creates new copies of all image copies of the database that have the tag full_cold_copy and gives the new copies the tag new_full_cold_copy.

Example 9-6 Assigning Tags to Output Copies

BACKUP AS COPY
  COPY OF DATABASE
    FROM TAG full_cold_copy
  TAG new_full_cold_copy;

Making Compressed Backups

For any use of the BACKUP command that creates backup sets, you can take advantage of RMAN support for binary compression of backup sets. Specify the AS COMPRESSED BACKUPSET option to the BACKUP command.

RMAN compresses the backup set contents before writing them to disk. The details of which binary compression level is used are automatically recorded in the backup set. There is no need to explicitly mention the type of compression used or how to decompress the backup set in the recovery operation.

Example 9-7 backs up the entire database and archived logs to the configured default backup destination (disk or tape), producing compressed backup sets.

Example 9-7 Making Compressed Backups

BACKUP 
  AS COMPRESSED BACKUPSET 
  DATABASE PLUS ARCHIVELOG;

Binary compression creates some performance overhead during backup and restore operations. Binary compression consumes CPU resources, so do not routinely schedule compressed backups when CPU usage is high. However, the following circumstances may warrant paying the performance penalty:

  • You are using disk-based backups when disk space in your fast recovery area or other disk-based backup destination is limited.

  • You are performing your backups to some device over a network when reduced network bandwidth is more important than CPU usage.

  • You are using some archival backup media such as CD or DVD, where reducing backup sizes saves on media costs and archival storage.

See Also:

"Binary Compression for Backup Sets" and the AS COMPRESSED BACKUPSET option of the BACKUP command in Oracle Database Backup and Recovery Reference for performance details regarding backup sets

Specifying Multisection Incremental Backups

A multisection backup enables large data files to be divided into sections that can be backed up in parallel across multiple channels. This provides faster backup performance and better recovery times. A multisection backup contains multiple backup pieces. During a multisection backup operation, RMAN writes to each backup piece, in parallel, by using a separate channel for each backup piece.

Multisection full backups of databases and data files are supported starting with Oracle Database 11g Release 1. Starting with Oracle Database 12c Release 1 (12.1), RMAN supports multisection incremental backups. Wherever applicable, RMAN also uses unused block compression and block change tracking while creating multisection incremental backups. When backup sets are used, you can create multisection full or incremental backups.

To create level 0 multisection incremental backups, the COMPATIBLE parameter must be set to 11.0 or higher. However, to create multisection incremental backups of level 1 or higher, you must set the COMPATIBLE parameter to 12.0.0 or higher. RMAN always creates multisection incremental backups with FILESPERSET set to 1.

Use the SECTION SIZE clause of the BACKUP command to create multisection backups. The SECTION SIZE clause specifies the size of each backup section. If you specify a section size that is larger than the size of the file, then RMAN does not use multisection backups for that file. If you specify a small section size that would produce more than 256 sections, then RMAN increases the section size to a value that results in exactly 256 sections.

Views to Identify Multisection Backups

Use the MULTI_SECTION column of the V$BACKUP_SET view or the recovery catalog view RC_BACKUP_SET to determine if a backup is a multisection backup. For multisection backups, the MULTI_SECTION column contains the value YES.

Views That Contain Metadata for Multisection Backups

The V$BACKUP_DATAFILE and RC_BACKUP_DATAFILE views provide information about the number of blocks in each section of a multisection backup. The BLOCKS column specifies the number of blocks in each multisection backup.

Example: Multisection Incremental Backup of Database as Backup Sets

Use the following steps to create a multisection level 1 incremental backup of a data file as backup sets:

  1. Ensure that the initialization parameter COMPATIBLE of the target database is set to 12.0.0 or higher.

  2. Start RMAN and connect to a target database as a user with the SYSBACKUP or SYSDBA privilege.

    See Also:

    "Making Database Connections with RMAN" for information about connecting to a target database
  3. If required, configure channel parallelism so that RMAN can write to the backup device in parallel.

    The following example configures two disk channels so that RMAN can back up to disk using two tape drives in parallel.

    CONFIGURE DEVICE TYPE disk PARALLELISM 2;
    
  4. Execute BACKUP with SECTION SIZE to indicate that a multisection backup must be created.

    The following example creates a multisection section level 1 backup of the data file users_df.dbf using backup sets. Each backup piece is 100MB.

    BACKUP 
       INCREMENTAL LEVEL 1
       SECTION SIZE 100M
       DATAFILE '/oradata/datafiles/users_df.dbf';
    

Making Multisection Backups Using Image Copies

RMAN enables you to create multisection backups using image copies. Multisection backups provide better performance by using multiple channels to back up large files in parallel. Starting with Oracle Database 12c Release 1 (12.1), you can create multisection full backups that are stored as image copies. While the image copy is being created, multiple channels are used to write files sections. However, the output of this operation is one copy for each data file.

Use the SECTION SIZE clause to create multisection backups. If the section size that you specify is larger than the size of the file, then RMAN does not use multisection backups for that file. If you specify a small section size that would produce more than 256 sections, then RMAN increases the section size to a value that results in exactly 256 sections

See "Specifying Multisection Incremental Backups" for information about the views that contain information about multisection backups.

Example: Multisection Backup of Data File as Image Copies

Use the following steps to create a multisection backup of a database as image copies:

  1. Ensure that the COMPATIBLE parameter for the target database is set to 12.0.0 or higher.

  2. Start RMAN and connect to a target database as a user with the SYSBACKUP or SYSDBA privilege.

    See Also:

    "Making Database Connections with RMAN" for information about connecting to a target database

    The sbu user is granted the SYSBACKUP privilege in the target database. RMAN prompts you to enter the password.

  3. If required, configure channel parallelism so that RMAN can perform the backup operation using multiple drives in parallel.

  4. Execute BACKUP with SECTION SIZE and AS COPY to indicate that a multisection backup must be created using image copies.

    The following example creates a multisection incremental backup of the entire database using image copies. Each backup piece is 500MB.

    BACKUP 
       AS COPY
       SECTION SIZE 500M
       DATABASE;
    

Backing Up Database Files with RMAN

This section contains the following topics:

Backing Up a Whole Database with RMAN

You can perform a whole database backup with the database mounted or open. To perform a whole database backup, from the RMAN prompt, use the BACKUP DATABASE command.

You may want to exclude specified tablespaces from a whole database backup. As explained in "Configuring Tablespaces for Exclusion from Whole Database Backups", you can persistently skip tablespaces across RMAN sessions by executing the CONFIGURE EXCLUDE command for each tablespace that you always want to skip. You can override the configured setting with BACKUP ... NOEXCLUDE.

To back up the database:

  1. Start RMAN and connect to a target database and a recovery catalog (if used).

    See Also:

    "Making Database Connections with RMAN" for information about connecting to a target database
  2. Ensure that the database is mounted or open.

  3. Issue the BACKUP DATABASE command at the RMAN prompt.

    The simplest form of the command requires no options or parameters:

    BACKUP DATABASE;
    

    For a list of what files are backed up (data files, control file, server parameter file) see the BACKUP command keyword DATABASE in Oracle Database Backup and Recovery Reference.

    The following example backs up the database, switches the online redo logs, and includes archived logs in the backup:

    BACKUP DATABASE PLUS ARCHIVELOG;
    

    By archiving the logs immediately after the backup, you ensure that you have a full set of archived logs through the time of the backup. In this way, you guarantee that you can perform media recovery after restoring this backup.

See Also:

Backing Up Tablespaces and Data Files with RMAN

You can back up one or more tablespaces with the BACKUP TABLESPACE command or one or more data files with the BACKUP DATAFILE command. When you specify tablespaces, RMAN translates the tablespace name internally into a list of data files. The database can be mounted or open. Tablespaces can be read/write or read-only.

Note:

Transportable tablespaces do not have to be in read/write mode for backup as in previous releases.

RMAN automatically backs up the control file and the server parameter file (if the instance was started with a server parameter file) when data file 1 is included in the backup. If control file autobackup is enabled, then RMAN writes the current control file and server parameter file to a separate autobackup piece. Otherwise, RMAN includes these files in the backup set that contains data file 1.

To back up tablespaces or data files:

  1. Start RMAN and connect to a target database and a recovery catalog (if used).

    See Also:

    "Making Database Connections with RMAN" for information about connecting to a target database
  2. If the database instance is not started, then either mount or open the database.

  3. Run the BACKUP TABLESPACE command or BACKUP DATAFILE command at the RMAN prompt.

The following example backs up the users and tools tablespaces to tape:

BACKUP
  DEVICE TYPE sbt
  TABLESPACE users, tools;

The following example uses an SBT channel to back up data files 1 through 4 and a data file copy stored at /tmp/system01.dbf to tape:

BACKUP 
  DEVICE TYPE sbt 
  DATAFILE 1,2,3,4 
  DATAFILECOPY '/tmp/system01.dbf';

Backing Up Control Files with RMAN

You can back up the control file when the database is mounted or open. RMAN uses a snapshot control file to ensure a read-consistent version. If the CONFIGURE CONTROLFILE AUTOBACKUP command is set to ON (by default it is OFF), then RMAN automatically backs up the control file and server parameter file after every backup and after database structural changes. The control file autobackup contains metadata about the previous backup, which is crucial for disaster recovery.

Note:

You can restore a backup of a control file made on one Data Guard database to any other database in the environment. Primary and standby control file backups are interchangeable. See Oracle Data Guard Concepts and Administration to learn how to use RMAN to restore files on a standby database.

If the autobackup feature is not set, then you must manually back up the control file in one of the following ways:

  • Run BACKUP CURRENT CONTROLFILE.

  • Include a backup of the control file within any backup by using the INCLUDE CURRENT CONTROLFILE option of the BACKUP command.

  • Back up data file 1, because RMAN automatically includes the control file and server parameter file in backups of data file 1.

    Note:

    If the control file block size is different from the block size for data file 1, then the control file cannot be written into the same backup set as the data file. RMAN writes the control file into a backup set by itself if the block size is different. The V$CONTROLFILE.BLOCK_SIZE column indicates the control file block size, whereas the DB_BLOCK_SIZE initialization parameter indicates the block size of data file 1.

About Manual Backups of the Control File

A manual backup of the control file is different from a control file autobackup. RMAN makes a control file autobackup after the files specified in the BACKUP command are backed up. Thus, the autobackup—unlike a manual control file backup—contains metadata about the backup that just completed. Also, RMAN can automatically restore autobackups without the use of a recovery catalog.

You can make a manual backup of the current control file either as a backup set or as an image copy. For a backup set, RMAN first creates a snapshot control file for read consistency. You can configure the file name and location of the snapshot control file. A snapshot control file is not needed for an image copy.

In an Oracle Real Application Clusters (Oracle RAC) environment, the following restrictions apply:

  • The snapshot control file location must be on shared storage—that is, storage that is accessible by all Oracle RAC instances.

  • The destination of an image copy of the current control file must be shared storage.

Making a Manual Backup of the Control File

To make a manual backup, you can either specify INCLUDE CURRENT CONTROLFILE when backing up other files or specify BACKUP CURRENT CONTROLFILE. You can also back up control file copies on disk by specifying the CONTROLFILECOPY parameter.

To manually back up the control file:

  1. Start RMAN and connect to a target database and a recovery catalog (if used).

    See Also:

    "Making Database Connections with RMAN" for information about connecting to a target database
  2. Ensure that the target database is mounted or open.

  3. Execute the BACKUP command with the desired control file clause.

    The following example backs up tablespace users to tape and includes the current control file in the backup:

    BACKUP DEVICE TYPE sbt 
      TABLESPACE users 
      INCLUDE CURRENT CONTROLFILE;
    

    The following example backs up the current control file to the fast recovery area as a backup set:

    BACKUP CURRENT CONTROLFILE;
    

    RMAN first creates a snapshot control file.

    The following example backs up the current control file to the default disk device as an image copy:

    BACKUP AS COPY
      CURRENT CONTROLFILE 
      FORMAT '/tmp/control01.ctl';
    

    The following example backs up the control file copy created in the previous example to tape:

    BACKUP AS COPY 
      CURRENT CONTROLFILE 
      FORMAT '/tmp/control01.ctl';
    BACKUP DEVICE TYPE sbt 
      CONTROLFILECOPY '/tmp/control01.ctl';
    

    A snapshot control file is not needed when backing up a control file copy.

    If the control file autobackup feature is enabled, then RMAN makes two control file backups in these examples: the explicit backup of the files specified in the BACKUP command and the control file and server parameter file autobackup.

    See Also:

    Oracle Database Backup and Recovery Reference to learn about the CONFIGURE CONTROLFILE AUTOBACKUP command

Backing Up Server Parameter Files with RMAN

As explained in "Backing Up Control Files with RMAN", RMAN automatically backs up the current server parameter file in certain cases. The BACKUP SPFILE command backs up the parameter file explicitly. The server parameter file that is backed up is the one currently in use by the instance.

To back up the server parameter file:

  1. Start RMAN and connect to a target database and a recovery catalog (if used).

    See Also:

    "Making Database Connections with RMAN" for information about connecting to a target database
  2. Ensure that the target database is mounted or open.

    The database must have been started with a server parameter file. If the instance is started with a client-side initialization parameter file, then RMAN issues an error if you execute BACKUP ... SPFILE.

  3. Execute the BACKUP ... SPFILE command.

    The following example backs up the server parameter file to tape:

    BACKUP DEVICE TYPE sbt SPFILE;
    

Backing Up a Database in NOARCHIVELOG Mode

You can only back up a database in NOARCHIVELOG mode when the database is closed and in a consistent state. The script shown in Example 9-8 puts the database into the correct mode for a consistent, whole database backup and then backs up the database. The script assumes that control file autobackup is enabled for the database.

Example 9-8 Backing Up a Database in NOARCHIVELOG Mode

SHUTDOWN IMMEDIATE; 
# Start the database in case it suffered instance failure or was 
# closed with SHUTDOWN ABORT before starting this script. 
STARTUP FORCE DBA; 
SHUTDOWN IMMEDIATE; 
STARTUP MOUNT;
# this example uses automatic channels to make the backup
BACKUP 
  INCREMENTAL LEVEL 0
  MAXSETSIZE 10M
  DATABASE
  TAG 'BACKUP_1';
# Now that the backup is complete, open the database. 
ALTER DATABASE OPEN; 

You can skip tablespaces, such as read-only tablespaces, but any skipped tablespace that has not been offline or read-only since its last backup is lost if the database has to be restored from a backup.

Backing Up CDBs and PDBs

This section explains how to back up multitenant container databases (CDBs) and pluggable databases (PDBs). It contains the following topics:

Note:

Backups of a non-CDB are not usable after the non-CDB is plugged in as a PDB into another CDB.

See Also:

Oracle Database Concepts for an introduction to PDBs

About Backing Up CDBs and PDBs

RMAN and Oracle Enterprise Manager Cloud Control provide full support for backup and recovery in a multitenant environment. The multitenant architecture enables an Oracle Database to function as a CDB. You can back up and recover a whole CDB, the root only, or one or more PDBs. You can also back up and recover individual tablespaces and data files in a PDB.

You might want to perform nightly backups of the whole CDB by using an incremental backup strategy, or you might want to make frequent separate backups of individual PDBs and do less frequent backups of either the whole CDB or of the root.

In terms of the ability to recover from data loss, separately backing up the root and all PDBs is equivalent to backing up the whole CDB. The main difference is in the number of RMAN commands that you must enter and the time to recover. Recovering a whole CDB requires less time than recovering the root plus all PDBs.

Backing Up a Whole CDB

Backing up a whole CDB is similar to backing up a non-CDB. When you back up a whole CDB, RMAN backs up the root, all the PDBs, and the archived redo logs. You can then recover either the whole CDB, the root only, or one or more PDBs from the CDB backup.

To back up a whole CDB:

Follow the instructions in "Backing Up a Whole Database with RMAN", connecting to the root as a common user with the SYSBACKUP or SYSDBA privilege.

Backing Up the Root with RMAN

You can use RMAN to make a backup of only the root. Because the root contains critical metadata for the whole CDB, Oracle recommends that you back up the root or back up the whole CDB at regular intervals.

To back up the root with RMAN:

  1. Start RMAN and connect to the root as a common user with the SYSBACKUP or SYSDBA privilege.

  2. Enter the following command:

    BACKUP DATABASE ROOT;
    

Backing Up the Root with Oracle Enterprise Manager Cloud Control

To back up the root with Oracle Enterprise Manager Cloud Control (Cloud Control), complete these steps:

  1. From the Database Home page, select Backup & Recovery from the Availability menu, and then select Schedule Backup.

  2. If you have not logged in to the database previously, then the Database Login page is displayed. Log in to the database using Named or New credentials and then click Login.

    Cloud Control displays the Schedule Backup page.

  3. From the Customized Backup section, choose Container Database Root, and then click Schedule Customized Backup.

    The Schedule Backup Wizard appears and displays the Options page.

  4. Complete the wizard by navigating the remainder of the pages to back up the root. For more information about each page of the wizard, click Help.

Backing Up PDBs with RMAN

RMAN enables you to back up one or more PDBs in a CDB. There are two approaches to backing up a PDB with RMAN:

  • Connect to the root and then use the BACKUP PLUGGABLE DATABASE command. This approach enables you to back up multiple PDBs with a single command.

    When you connect to the root and back up a PDB, this backup is visible to the root and to that particular PDB but not to the other PDBs.

  • Connect to the PDB and use the BACKUP DATABASE command. This approach backs up only a single PDB and enables you to use the same commands used for backing up non-CDBs.

    Backups created when connected to any PDB are visible when connected to the root.

When you back up individual PDBs, the archived redo logs are not backed up.

To back up one or more PDBs while connected to the root:

  1. Start RMAN and connect to the root as a common user with the SYSBACKUP or SYSDBA privilege.

  2. Issue a BACKUP PLUGGABLE DATABASE command at the RMAN prompt.

    The following example backs up the PDBs sales and hr:

    BACKUP PLUGGABLE DATABASE sales, hr;
    

To back up one PDB while connected to the PDB:

  1. Start RMAN and connect to the PDB as a local user with the SYSBACKUP or SYSDBA privilege.

  2. Issue a BACKUP DATABASE command at the RMAN prompt.

    BACKUP DATABASE;
    

Backing Up PDBs with Oracle Enterprise Manager Cloud Control

To back up one or more PDBs with Oracle Enterprise Manager Cloud Control (Cloud Control), complete these steps:

  1. From the Database Home page, select Backup & Recovery from the Availability menu, and then select Schedule Backup.

  2. If you have not logged in to the database previously, then the Database Login page is displayed. Log in to the database using Named or New credentials and then click Login.

    Cloud Control displays the Schedule Backup page.

  3. From the Customized Backup section, select Pluggable Databases, and then click Schedule Customized Backup.

    The Schedule Backup Wizard appears and displays the Pluggable Databases page.

  4. Select the PDBs that you want to back up by following these steps:

    1. Click Add to display the Available Pluggable Databases page.

    2. From the list of PDBs shown, click in the Select column to designate the PDBs you want to back up. Optionally, you can click Select All to turn on the Select option for all available PDBs. Click Select None to deselect all PDBs.

    3. Click the Select button to return to the Pluggable Databases page.

    4. Optionally, you can remove PDBs from the table by clicking in the Select column for each PDB that you want to remove and then clicking Remove.

  5. Click Next to move to the Options page of the wizard.

  6. Complete the wizard by navigating the remainder of the pages to back up the PDBs. For more information about each page of the wizard, click Help.

Backing Up Tablespaces and Data Files in a PDB

Because tablespaces in different PDBs can have the same name, to eliminate ambiguity you must connect directly to a PDB to back up one or more of its tablespaces. In contrast, because data file numbers and paths are unique across the CDB, you can connect to either the root or a PDB to back up PDB data files. If you connect to the root, you can back up data files from multiple PDBs with a single command. If you connect to a PDB, you can back up only data files in that PDB.

To back up tablespaces in a PDB:

  1. Start RMAN and connect to the PDB as a local user with the SYSBACKUP or SYSDBA privilege.

  2. Issue a BACKUP TABLESPACE command as described in "Backing Up Tablespaces and Data Files with RMAN".

    BACKUP TABLESPACE users, examples;
    

To back up data files in a PDB:

  1. Do one of the following:

    • Start RMAN and connect to the root as a common user with the SYSBACKUP or SYSDBA privilege.

    • Start RMAN and connect to the PDB as a local user with the SYSBACKUP or SYSDBA privilege.

  2. Issue a BACKUP DATAFILE command.

    BACKUP DATAFILE 10, 13;
    

Backing Up Archived Redo Logs with RMAN

Archived redo logs are the key to successful media recovery. You should back them up regularly.

This section contains the following topics:

About Backups of Archived Redo Logs for non-CDBs

Several features of RMAN backups are specific to archived redo logs. For example, you can use BACKUP ... DELETE to delete one or all copies of archived redo logs from disk after backing them up to backup sets.

Archived Redo Log Failover

Even if your redo logs are being archived to multiple destinations and you use RMAN to back up archived redo logs, RMAN selects only one copy of the archived redo log file to include in the backup set. Because logs with the same log sequence number are identical, RMAN does not need to include more than one log copy.

The archived redo log failover feature enables RMAN to complete a backup even when some archiving destinations are missing logs or contain logs with corrupt blocks. If at least one log corresponding to a given log sequence and thread is available in the fast recovery area or any of the archiving destinations, then RMAN tries to back it up. If RMAN finds a corrupt block in a log file during backup, it searches other destinations for a copy of that log without corrupt blocks.

For example, assume that you archive logs 121 through 124 to two destinations: /arch1 and /arch2. Table 9-1 shows the archived redo log records in the control file.

Table 9-1 Sample Archived Redo Log Records

Sequence File Name in /arch1 File Name in /arch2

121

/arch1/archive1_121.arc

/arch2/archive1_121.arc

122

/arch1/archive1_122.arc

/arch2/archive1_122.arc

123

/arch1/archive1_123.arc

/arch2/archive1_123.arc

124

/arch1/archive1_124.arc

/arch2/archive1_124.arc


However, unknown to RMAN, a user deletes logs 122 and 124 from the /arch1 directory. Afterward, you run the following backup:

BACKUP ARCHIVELOG
  FROM  SEQUENCE 121 
  UNTIL SEQUENCE 125;

With failover, RMAN completes the backup, using logs 122 and 124 in /arch2.

Online Redo Log Switching

Another important RMAN feature is automatic online redo log switching. To make an open database backup of archived redo logs that includes the most recent online redo log, you can execute the BACKUP command with any of the following clauses:

  • PLUS ARCHIVELOG

  • ARCHIVELOG ALL

  • ARCHIVELOG FROM ...

Before beginning the backup, RMAN switches out of the current redo log group, and archives all online redo logs that have not yet been archived, up to and including the redo log group that was current when the command was issued. This feature ensures that the backup contains all redo generated before the start of the command.

An effective way of backing up archived redo logs is the BACKUP ... PLUS ARCHIVELOG command, which causes RMAN to do the following:

  1. Run the ALTER SYSTEM ARCHIVE LOG CURRENT statement.

  2. Run BACKUP ARCHIVELOG ALL. If backup optimization is enabled, then RMAN skips logs that it has already backed up to the specified device.

  3. Back up the rest of the files specified in the BACKUP command.

  4. Run the ALTER SYSTEM ARCHIVE LOG CURRENT statement.

  5. Back up any remaining archived logs generated during the backup. If backup optimization is not enabled, then RMAN backs up the logs generated in Step 1 plus all the logs generated during the backup.

The preceding steps guarantee that data file backups taken during the command are recoverable to a consistent state. Also, unless the online redo log is archived at the end of the backup, DUPLICATE is not possible with the backup.

About Backup of Archived Redo Logs in CDBs

In a CDB, archived redo logs can be backed up only when you connect to the root as a common user with the SYSDBA or SYSBACKUP privilege. When you connect to a PDB as a local user with SYSDBA or SYSBACKUP privilege, you cannot back up or delete archived redo logs.

If your archived redo logs are being copied to multiple destinations, when you connect to the root and backup archived redo log files, RMAN includes only one copy of the archived redo log files in a backup. You can switch archived redo log files when you connect to root of a CDB. Therefore, the information in "Archived Redo Log Failover" and "Online Redo Log Switching" is applicable when you connect to the root. However, you cannot backup or switch archived redo log files when connected to a PDB.

Backing Up Archived Redo Log Files in non-CDBs

To back up archived logs, use the BACKUP ARCHIVELOG command. If backup optimization is enabled, then RMAN skips backups of archived logs that have already been backed up to the specified device.

To back up archived redo log files:

  1. Start RMAN and connect to a target database and a recovery catalog (if used).

    See Also:

    "Making Database Connections with RMAN" for information about connecting to a target database
  2. Ensure that the target database is mounted or open.

  3. Execute the BACKUP ARCHIVELOG or BACKUP ... PLUS ARCHIVELOG command.

    The following example backs up the database and all archived redo logs:

    BACKUP DATABASE PLUS ARCHIVELOG;
    

    The following example uses a configured disk or SBT channel to back up one copy of each log sequence number for all archived redo logs:

    BACKUP ARCHIVELOG ALL;
    

    You can also specify a range of archived redo logs by time, SCN, or log sequence number, as in the following example:

    BACKUP ARCHIVELOG 
      FROM TIME  'SYSDATE-30'
      UNTIL TIME 'SYSDATE-7';
    

Backing Up Only Archived Redo Logs That Need Backups in non-CDBs

You can indicate that RMAN should automatically skip backups of archived redo logs in the following ways:

The BACKUP ... NOT BACKED UP integer TIMES command specifies that RMAN backs up only those archived log files that have not been backed up at least integer times to the specified device. To determine the number of backups for a file, RMAN only considers backups created on the same device type as the current backup.

The BACKED UP clause is a convenient way to back up archived logs to a specified device type. For example, you can specify that RMAN should keep two copies of each archived redo log on tape and skip additional backups.

To back up archived redo logs that need backups:

  1. Start RMAN and connect to a target database and a recovery catalog (if used).

    See Also:

    "Making Database Connections with RMAN" for information about connecting to a target database
  2. Ensure that the target database is mounted or open.

  3. Ensure that appropriate channels are configured for the backup.

  4. Execute the BACKUP ARCHIVELOG command with the NOT BACKED UP clause.

    BACKUP ARCHIVELOG ALL NOT BACKED UP 2 TIMES;
    

See Also:

"Using Backup Optimization to Skip Files" for scenarios using NOT BACKED UP

Backing Up Archived Redo Logs in CDBs

You can back up archived redo logs in a CDB by using the BACKUP ARCHIVELOG command.

To back up archived redo logs in a CDB:

Start RMAN and connect to the root as a user with the SYSDBA or SYSBACKUP privilege. Follow the instructions in "Backing Up Archived Redo Log Files in non-CDBs".

To back up only archived redo logs that need backup in a CDB:

Start RMAN and connect to the root as a common user with the SYSDBA or SYSBACKUP privilege. Follow the instructions in "Backing Up Only Archived Redo Logs That Need Backups in non-CDBs".

Deleting Archived Redo Logs After Backups in non-CDBs

The BACKUP ARCHIVELOG ... DELETE INPUT command deletes archived log files after they are backed up. This command eliminates the separate step of manually deleting archived redo logs.

With DELETE INPUT, RMAN deletes only the specific copy of the archived log chosen for the backup set. With DELETE ALL INPUT, RMAN deletes each backed-up archived redo log file from all log archiving destinations.

As explained in "Configuring an Archived Redo Log Deletion Policy", the BACKUP ... DELETE INPUT and DELETE ARCHIVELOG commands obey the archived redo log deletion policy for logs in all archiving locations. For example, if you specify that logs be deleted only when backed up at least twice to tape, then BACKUP ... DELETE honors this policy.

For the following procedure, assume that you archive to /arc_dest1, /arc_dest2, and the fast recovery area.

To delete archived redo logs after a backup:

  1. Start RMAN and connect to a target database and a recovery catalog (if used).

    See Also:

    "Making Database Connections with RMAN" for information about connecting to a target database
  2. Ensure that the target database is mounted or open.

  3. Run the BACKUP command with the DELETE INPUT clause.

    Assume that you run the following BACKUP command:

    BACKUP DEVICE TYPE sbt 
      ARCHIVELOG ALL 
        DELETE ALL INPUT;
    

    In this case, RMAN backs up only one copy of each log sequence number in these archiving locations. RMAN deletes all copies of any log that it backed up from both the fast recovery area and the other archiving destinations.

    If you specify DELETE INPUT rather than DELETE ALL INPUT, then RMAN only deletes the specific archived redo log files that it backed up. For example, RMAN deletes the logs in /arc_dest1 if these files were used as the source of the backup, but leave the contents of the /arc_dest2 intact.

See Also:

Deleting Archived Redo Logs After Backups in CDBs

In a CDB, you can delete archived redo logs after they are backed up by using the BACKUP ARCHIVELOG ... DELETE INPUT command.

To delete archived redo logs in a CDB after they are backed up:

Start RMAN and connect to the root as a user with the SYSDBA or SYSBACKUP privilege. Follow the instructions in "Deleting Archived Redo Logs After Backups in non-CDBs".

Making and Updating Incremental Backups

As explained in "Incremental Backups", an incremental backup copies only data file blocks that have changed since a specified previous backup. An incremental backup is either a cumulative incremental backup or a differential incremental backup.

Although the content of the backups is the same, BACKUP DATABASE and BACKUP INCREMENTAL LEVEL 0 DATABASE are different. A full backup is not usable as part of an incremental strategy, whereas a level 0 incremental backup is the basis of an incremental strategy. No RMAN command can change a full backup into a level 0 incremental backup.

As with full backups, RMAN can make incremental backups of an ARCHIVELOG mode database that is open. If the database is in NOARCHIVELOG mode, then RMAN can make incremental backups only after a consistent shutdown.

Purpose of Incremental Backups

The primary reasons for making incremental backups part of your strategy are:

  • Faster daily backups if block change tracking is enabled (see "Using Block Change Tracking to Improve Incremental Backup Performance")

  • Ability to roll forward data file image copies, thereby reducing recovery time and avoiding repeated full backups

  • Less bandwidth consumption when backing up over a network

  • Improved performance when the aggregate tape bandwidth for tape write I/Os is much less than the aggregate disk bandwidth for disk read I/Os

  • Possibility of recovering changes to objects created with the NOLOGGING option

    For example, direct load inserts do not create redo log entries, so their changes cannot be reproduced with media recovery. Direct load inserts do change data blocks, however, and these blocks are captured by incremental backups.

  • Ability to synchronize a physical standby database with the primary database

    You can use the RMAN BACKUP INCREMENTAL FROM SCN command to create a backup on the primary database that starts at the current SCN of the standby database, which you can then use to roll forward the standby database. See Oracle Data Guard Concepts and Administration to learn how to apply incremental backups to a standby database.

See Also:

Oracle Database Administrator's Guide for more information about NOLOGGING mode

Planning an Incremental Backup Strategy

Choose a backup strategy according to an acceptable MTTR (mean time to recover). For example, you can implement a three-level backup scheme so that a level 0 backup is taken monthly, a cumulative level 1 is taken weekly, and a differential level 1 is taken daily. In this strategy, you never have to apply more than a day of redo for complete recovery.

When deciding how often to take level 0 backups, a general rule is to take a new level 0 backup whenever 20% or more of the data has changed. If the rate of change to your database is predictable, then you can observe the size of your incremental backups to determine when a new level 0 backup is appropriate. The following SQL query determines the number of blocks written to an incremental level 1 backup of each data file with at least 20% of its blocks backed up:

SELECT   FILE#, INCREMENTAL_LEVEL, COMPLETION_TIME, 
         BLOCKS, DATAFILE_BLOCKS 
FROM     V$BACKUP_DATAFILE 
WHERE    INCREMENTAL_LEVEL > 0 
AND      BLOCKS / DATAFILE_BLOCKS > .2
ORDER BY COMPLETION_TIME;

Compare the number of blocks in level 1 backups to a level 0 backup. For example, if you create only level 1 cumulative backups, then take a new level 0 backup when the most recent level 1 backup is about half the size of the level 0 backup.

An effective way to conserve disk space is to make incremental backups to disk, and then offload the backups to tape with the BACKUP AS BACKUPSET command. Incremental backups are generally smaller than full backups, which limits the space required to store them until they are moved to tape. When the incremental backups on disk are backed up to tape, the tape is more likely to stream because all blocks of the incremental backup are copied to tape. There is no possibility of delay due to time required for RMAN to locate changed blocks in the data files.

Another strategy is to use incrementally updated backups, as explained in "Incrementally Updating Backups". In this strategy, you create an image copy of each data file, and then periodically roll forward this copy by making and then applying a level 1 incremental backup. In this way you avoid the overhead of making repeated full image copies of your data files, but enjoy all of the advantages.

In a Data Guard environment, you can offload incremental backups to a physical standby database. Incremental backups of a standby and primary database are interchangeable. Thus, you can apply an incremental backup of a standby database to a primary database, or apply an incremental backup of a primary database to a standby database.

See Also:

Oracle Data Guard Concepts and Administration to learn how to back up a standby database with RMAN. In particular, consult Chapter 10, "Managing Physical and Snapshot Standby Databases"

Making Incremental Backups

After starting RMAN, run the BACKUP INCREMENTAL command at the RMAN prompt. By default incremental backups are differential.

To make an incremental backup:

  1. Start RMAN and connect to a target database and a recovery catalog (if used).

  2. Ensure that the target database is mounted or open.

  3. Execute the BACKUP INCREMENTAL command with the desired options.

    Use the LEVEL parameter to indicate the incremental level. The following example makes a level 0 incremental database backup.

    BACKUP
      INCREMENTAL LEVEL 0
      DATABASE;
    

    The following example makes a differential incremental backup at level 1 of the SYSTEM and tools tablespaces. It only backs up those data blocks changed since the most recent level 1 or level 0 backup.

    BACKUP
      INCREMENTAL LEVEL 1
      TABLESPACE SYSTEM, tools;
    

    The following example makes a cumulative incremental backup at level 1 of the tablespace users, backing up all blocks changed since the most recent level 0 backup.

    BACKUP
      INCREMENTAL LEVEL 1 CUMULATIVE
      TABLESPACE users;
    

Making Incremental Backups of a VSS Snapshot

You can use the Volume Shadow Copy Service (VSS) with the Oracle VSS writer to make a shadow copy or snapshot of files in a database. You must use a third-party backup program other than RMAN to make VSS snapshots with the Oracle VSS writer. In this case, the fast recovery area automates management of files that are backed up in a VSS snapshot and deletes them as needed.

You can use the BACKUP INCREMENTAL LEVEL 1 ... FROM SCN command in RMAN to create incremental backups in the fast recovery area. Thus, you can use this command to create an incremental level 1 backup of a VSS shadow copy. RMAN can apply incremental backups during recovery transparently.

See Also:

Oracle Database Platform Guide for Microsoft Windows to learn how to make VSS backups with RMAN

Incrementally Updating Backups

By incrementally updating backups, you can avoid the overhead of making full image copy backups of data files, while also minimizing time required for media recovery of your database. For example, if you run a daily backup script, then you never have more than 1 day of redo to apply for media recovery.

To incrementally update data file backups:

  1. Create a full image copy backup of a data file with a specified tag.

  2. At regular intervals (such as daily), make a level 1 differential incremental backup of the data file and use the same tag as the base data file copy.

  3. Apply the incremental backup to the most recent backup with the same tag.

This technique rolls forward the backup to the time when the level 1 incremental backup was made. RMAN can restore this incremental forever and apply changes from the redo log. The result equals restoring a data file backup taken at the SCN of the most recently applied incremental level 1 backup.

Note:

If you run RECOVER COPY daily without specifying an UNTIL TIME, then a continuously updated image copy cannot satisfy a recovery window of more than a day. The incrementally updated backup feature is an optimization for fast media recovery.

Incrementally Updating Backups: Basic Example

To create incremental backups for use in an incrementally updated backup strategy, use the BACKUP ... FOR RECOVER OF COPY WITH TAG form of the BACKUP command. The command is best understood in a sample script that implements the strategy.

The script in Example 9-9, run regularly, is all that is required to implement a strategy based on incrementally updated backups.

Example 9-9 Basic Incremental Update Script

RUN
{
  RECOVER COPY OF DATABASE 
    WITH TAG 'incr_update';
  BACKUP 
    INCREMENTAL LEVEL 1
    FOR RECOVER OF COPY WITH TAG 'incr_update'
    DATABASE;
}

To understand the script and the strategy, you must understand the effects of these two commands when no data file copies or incremental backups exist. Note two important features:

  • The BACKUP command in Example 9-9 does not always create a level 1 incremental backup.

  • The RECOVER command in Example 9-9 causes RMAN to apply any available incremental level 1 backups with the specified tag to a set of data file copies with the same tag.

Table 9-2 shows the effect of the script when it is run once per day starting on Monday.

Table 9-2 Effect of Basic Script When Run Daily

Command Monday Tuesday Wednesday Thursday Onward

RECOVER

Because no incremental backup or data file copy exists, the command generates a message (but not an error). That is, the command has no effect.

A database copy now exists, but no incremental level 1 backup exists with which to recover it. Thus, the RECOVER command has no effect.

The level 1 incremental backup made on Tuesday is applied to the database copy, bringing the copy up to the checkpoint SCN of the level 1 incremental backup.

The level 1 incremental backup made yesterday is applied to the database copy, bringing the copy up to the checkpoint SCN of the level 1 incremental backup.

BACKUP

No level 0 image copy exists, so the command creates an image copy of the database and applies the tag incr_update. This copy is needed to begin the cycle of incremental updates.

Note: If the script sets DEVICE TYPE sbt, then the first run creates the copy on disk, not on tape. Subsequent runs make level 1 backups on tape.

The command makes an incremental level 1 backup and assigns it the tag incr_update. This backup contains blocks that changed between Monday and Tuesday.

The command makes an incremental level 1 backup and assigns it the tag incr_update. This backup contains blocks that changed between Tuesday and Wednesday.

The command makes an incremental level 1 backup and assigns it the tag incr_update. This backup contains blocks that changed between now and the most recent backup with the tag incr_update.


Note the following additional details about Example 9-9:

  • Each time a data file is added to the database, an image copy of the new data file is created the next time the script runs. The next run makes the first level 1 incremental for the added data file. On all subsequent runs the new data file is processed like any other data file.

  • You must use tags to identify the data file copies and incremental backups in this strategy so that they do not interfere with other backup strategies. If you use multiple incremental backup strategies, then RMAN cannot unambiguously create incremental level 1 backups unless you tag level 0 backups.

    The incremental level 1 backups to apply to those image copies are selected based upon the tag of the image copy data files and the available incremental level 1 backups. The tag is essential in the selection of the incremental level backups.

  • After the third run of the script, the following files are available for a point-in-time recovery:

    • An image copy of the database, as of the checkpoint SCN of the preceding run of the script, 24 hours earlier

    • An incremental backup for the changes after the checkpoint SCN of the preceding run

    • Archived redo logs including all changes between the checkpoint SCN of the image copy and the current time

    If you must restore and recover your database during the following 24 hours, then you can restore the data files from the incrementally updated data file copies. You can then apply changes from the most recent incremental level 1 and the redo logs to reach the desired SCN. At most, you have 24 hours of redo to apply, which limits how long point-in-time recovery takes to finish.

Incrementally Updated Backups: Advanced Example

You can extend the basic script in Example 9-9 to provide fast recoverability to a window greater than 24 hours. Example 9-10 shows how to maintain a window of 7 days by specifying the beginning time of your window of recoverability in the RECOVER command.

Example 9-10 Advanced Incremental Update Script

RUN
{
  RECOVER COPY OF DATABASE 
    WITH TAG 'incr_update' 
    UNTIL TIME 'SYSDATE - 7';
  BACKUP
    INCREMENTAL LEVEL 1 
    FOR RECOVER OF COPY WITH TAG 'incr_update'
    DATABASE;
}

Table 9-3 shows the effect of the script when it is run once per day starting on Monday, January 1.

Table 9-3 Effect of Advanced Script When Run Daily

Command Monday 1/1 Tuesday 1/2 - Monday 1/8 Tuesday 1/9 Wednesday 1/10 Onward

RECOVER

Because no incremental backup or data file copy exists, the command generates a message (but not an error). That is, the command has no effect.

A database copy exists, but SYSDATE-7 specifies a time before the base copy was created. For example, on Wednesday SYSDATE-7 specifies the Wednesday before Monday 1/1. Thus, the RECOVER command has no effect.

SYSDATE-7 now specifies a date after the base copy was created. The database copy made on Monday 1/1 is updated with the incremental backup made on Tuesday 1/2, bringing the copy up to the checkpoint SCN of the level 1 incremental backup.

The database copy is updated with the incremental backup made 7 days ago, bringing the copy up to the checkpoint SCN of the level 1 incremental backup.

BACKUP

No level 0 image copy exists, so the command creates an image copy of the database and applies the tag incr_update. This copy is needed to begin the cycle of incremental updates.

Note: If the script sets DEVICE TYPE sbt, then the first run creates the copy on disk, not on tape. Subsequent runs make level 1 backups on tape.

The command makes an incremental level 1 backup and assigns it the tag incr_update. This backup contains blocks that changed between yesterday and today.

The command makes an incremental level 1 backup and assigns it the tag incr_update. This backup contains blocks that changed between Monday 1/8 and Tuesday 1/9.

The command makes an incremental level 1 backup and assigns it the tag incr_update. This backup contains blocks that changed between yesterday and today.


As with the basic script in Example 9-9, you have fast recoverability to any point in time between the SCN of the data file copies and the present. RMAN can use both block changes from the incremental backups and individual changes from the redo logs. Because you have the daily level 1 incremental backups, you never need to apply more than 1 day of redo.

See Also:

Oracle Database Backup and Recovery Reference to learn about the RECOVER command

Using Block Change Tracking to Improve Incremental Backup Performance

The block change tracking feature for incremental backups improves backup performance by recording changed blocks for each data file.

About Block Change Tracking

If block change tracking is enabled on a primary or standby database, then RMAN uses a block change tracking file to identify changed blocks for incremental backups. By reading this small bitmap file to determine which blocks changed, RMAN avoids having to scan every block in the data file that it is backing up.

Block change tracking is disabled by default. Nevertheless, the benefits of avoiding full data file scans during backup are considerable, especially if only a small percentage of data blocks are changed between backups. If your backup strategy involves incremental backups, then block change tracking is recommended. Block change tracking does not change the commands used to perform incremental backups. The change tracking file requires no maintenance after initial configuration.

You can only enable block change tracking at a physical standby database if a license for the Oracle Active Data Guard option is enabled.

Space Management in the Block Change Tracking File

The change tracking file maintains bitmaps that mark changes in the data files between backups. The database performs a bitmap switch before each backup. Oracle Database automatically manages space in the change tracking file to retain block change data that covers the eight most recent backups. After the maximum of eight bitmaps is reached, the oldest bitmap is overwritten by the bitmap that tracks the current changes.

The first level 0 incremental backup scans the entire data file. Subsequent incremental backups use the block change tracking file to scan only the blocks that have been marked as changed since the last backup. An incremental backup can be optimized only when it is based on a parent backup that was made after the start of the oldest bitmap in the block change tracking file.

Consider the eight-bitmap limit when developing your incremental backup strategy. For example, if you make a level 0 database backup followed by seven differential incremental backups, then the block change tracking file now includes eight bitmaps. If you then make a cumulative level 1 incremental backup, then RMAN cannot optimize the backup, because the bitmap corresponding to the parent level 0 backup is overwritten with the bitmap that tracks the current changes.

Location of the Block Change Tracking File

One block change tracking file is created for the whole database. By default, the block change tracking file is created as an Oracle managed file in the destination specified by the DB_CREATE_FILE_DEST initialization parameter. You can also place the block change tracking file in any location that you choose, by specifying its name when enabling block change tracking. Oracle recommends against using a raw device (that is, a disk without a file system) as a change tracking file.

Note:

In an Oracle RAC environment, the change tracking file must be located on shared storage accessible from all nodes in the cluster.

RMAN does not support backup and recovery of the change tracking file. The database resets the change tracking file when it determines that the change tracking file is invalid. If you restore and recover the whole database or a subset, then the database resets the block change tracking file and starts tracking changes again. After you make a level 0 incremental backup, the next incremental backup can use change tracking data.

Size of the Block Change Tracking File

The size of the block change tracking file is proportional to the size of the database and the number of enabled threads of redo. The size of the block change tracking file can increase and decrease as the database changes. The size is not related to the frequency of updates to the database.

Typically, the space required for block change tracking for a single instance is approximately 1/30,000 the size of the data blocks to be tracked. For an Oracle RAC environment, it is 1/30,000 of the size of the database, times the number of enabled threads.

The following factors that may cause the file to be larger than this estimate suggests:

  • To avoid the overhead of allocating space as your database grows, the block change tracking file size starts at 10 megabytes. New space is allocated in 10 MB increments. Thus, for any database up to approximately 300 gigabytes, the file size is no smaller than 10 MB, for up to approximately 600 gigabytes the file size is no smaller than 20 megabytes, and so on.

  • For each data file, a minimum of 320 kilobytes of space is allocated in the block change tracking file, regardless of the size of the data file. Thus, if you have a large number of relatively small data files, the change tracking file is larger than for databases with a smaller number of larger data files containing the same data.

Enabling and Disabling Block Change Tracking

You can enable block change tracking when the database is either open or mounted. This section assumes that you intend to create the block change tracking file as an Oracle managed file in the database area, which is where the database maintains active database files such as data files, control files, and online redo log files. See "Overview of the Fast Recovery Area" to learn about the database area and fast recovery area.

To enable block change tracking:

  1. Start SQL*Plus and connect to a target database with administrator privileges.

  2. Ensure that the DB_CREATE_FILE_DEST initialization parameter is set.

    SHOW PARAMETER DB_CREATE_FILE_DEST
    

    If the parameter is not set, and if the database is open, then you can set the parameter with the following form of the ALTER SYSTEM statement:

    ALTER SYSTEM SET 
      DB_CREATE_FILE_DEST = '/disk1/bct/'
      SCOPE=BOTH SID='*';
    
  3. Enable block change tracking.

    Execute the following ALTER DATABASE statement:

    ALTER DATABASE ENABLE BLOCK CHANGE TRACKING;
    

    You can also create the change tracking file in a location that you choose yourself by using the following form of SQL statement:

    ALTER DATABASE ENABLE BLOCK CHANGE TRACKING 
      USING FILE '/mydir/rman_change_track.f' REUSE;
    

    The REUSE option tells Oracle Database to overwrite any existing block change tracking file with the specified name.

Disabling Block Change Tracking

This section assumes that the block change tracking feature is currently enabled. When you disable block change tracking, the database removes the block change tracking file from the operating system.

To disable block change tracking:

  1. Start SQL*Plus and connect to a target database with administrator privileges.

  2. Ensure that the target database is mounted or open.

  3. Disable block change tracking.

    Execute the following ALTER DATABASE statement:

    ALTER DATABASE DISABLE BLOCK CHANGE TRACKING;
    

Checking Whether Change Tracking Is Enabled

You can query the V$BLOCK_CHANGE_TRACKING view to determine whether change tracking is enabled, and if it is, the file name of the block change tracking file.

To determine whether change tracking is enabled:

Enter the following query in SQL*Plus (sample output included):

COL STATUS   FORMAT A8
COL FILENAME FORMAT A60
 
SELECT STATUS, FILENAME
FROM   V$BLOCK_CHANGE_TRACKING;

STATUS   FILENAME
-------- ------------------------------------------------------------
ENABLED  /disk1/bct/RDBMS/changetracking/o1_mf_2f71np5j_.chg

Changing the Location of the Block Change Tracking File

To move the change tracking file, use the ALTER DATABASE RENAME FILE statement. The database must be mounted. The statement updates the control file to refer to the new location and preserves the contents of the change tracking file. If you cannot shut down the database, then you can disable and enable block change tracking. In this case, you lose the contents of the existing block change tracking file.

To change the location of the change tracking file:

  1. Start SQL*Plus and connect to a target database.

  2. If necessary, determine the current name of the change tracking file:

    SQL> SELECT FILENAME FROM V$BLOCK_CHANGE_TRACKING;
    
  3. If possible, shut down the database. For example:

    SQL> SHUTDOWN IMMEDIATE
    

    If you shut down the database, then skip to the next step. If you choose not to shut down the database, then execute the following SQL statements and skip all remaining steps:

    SQL> ALTER DATABASE DISABLE BLOCK CHANGE TRACKING;
    SQL> ALTER DATABASE ENABLE BLOCK CHANGE TRACKING USING FILE 'new_location';
    

    In this case you lose the contents of the block change tracking file. Until the next time you complete a level 0 incremental backup, RMAN must scan the entire file.

  4. Using host operating system commands, move the block change tracking file to its new location.

  5. Mount the database and move the change tracking file to a location that has more space. For example:

    ALTER DATABASE RENAME FILE
       '/disk1/bct/RDBMS/changetracking/o1_mf_2f71np5j_.chg' TO 
       '/disk2/bct/RDBMS/changetracking/o1_mf_2f71np5j_.chg';
    

    This statement changes the location of the change tracking file while preserving its contents.

  6. Open the database:

    SQL> ALTER DATABASE OPEN;
    

See Also:

Oracle Database SQL Language Reference to learn about the ALTER DATABASE statement and the ALTER SYSTEM statement

Making Database Backups for Long-Term Storage

This section explains the basic concepts and tasks involved in making backups for long-term storage.

Purpose of Archival Backups

You can use BACKUP ... KEEP to create a backup that is both all-inclusive and exempt from the backup retention policy. The backup is all-inclusive because every file needed to restore and recover the database is backed up to a single disk or tape location. The KEEP option also specifies that the backup is exempt from the retention policy either forever or for a specified period. The general name for a backup created with BACKUP ... KEEP is an archival backup.

As explained in "Data Archival", one purpose of a backup and recovery strategy is to preserve data. You can use BACKUP ... KEEP to retain a database backup for longer than the time dictated by the retention policy. For example, you can back up the database on the first day of every year to satisfy a regulatory requirement and store the media off-site. Years after you make the archival backup, you can restore and recover it to query the data as it appeared at the time of the backup.

Another purpose of an archival backup is to create a backup that you want to restore for testing purposes and then delete. For example, you can back up the database, restore the database in a test environment, and then discard the archival backup after the test database is operational. A related purpose is to create a self-contained backup that you can delete after transferring it to another user or host. For example, another user might want a copy of the database for reporting or testing.

Basic Concepts of Archival Backups

You can exempt a backup from the retention policy by using the KEEP option with the BACKUP command. You can also use the KEEP and NOKEEP options of the CHANGE command to change the status of an existing backup. Backups with KEEP attributes are valid backups that can be recovered like any other backups.

You can specify an end date for an archival backup with the KEEP UNTIL TIME clause, or specify that the backup is kept FOREVER. If you specify UNTIL, then RMAN marks the backup as obsolete when the UNTIL time has passed, regardless of any configured retention policy. For example, if you specify KEEP UNTIL TIME '01-JAN-13', then the backup is obsolete one second after midnight on January 1, 2013. If you specify an UNTIL TIME of 9:00 p.m, then the backup is obsolete at 9:01 p.m.

When you specify KEEP on the BACKUP command, RMAN generates multiple backup sets. Note the following characteristics of the BACKUP ... KEEP command:

  • It automatically backs up the data files, control file (even if the control file autobackup is disabled), and the server parameter file.

  • It automatically generates an archived redo log backup to ensure that the database backup can be recovered to a consistent state.

  • If the FORMAT, POOL, or TAG parameters are specified, then they are used for all backups. For this reason, the FORMAT string must allow for the creation of multiple backup pieces. Specifying the %U substitution variable is the easiest way to meet this requirement.

  • It supports an optional RESTORE POINT clause that creates a normal restore point, which is a label for an SCN to which the backup must be recovered to be made consistent. The SCN is captured just after the data file backups complete. RMAN resynchronizes restore points with the recovery catalog and maintains the restore points while the backup exists. "Listing Restore Points" explains how to display restore points.

See Also:

Oracle Database Backup and Recovery Reference for CHANGE syntax and Oracle Database Backup and Recovery Reference for BACKUP ... KEEP syntax

Making an Archival Backup for Long-Term Storage

Typically, you make an archival backup to tape. Because your data protection backups are most likely to be on a set of tapes that remain accessible and are recycled, it is advisable to reserve a set of tapes for the archival backup. You can write the archival backup to this special set of tapes and then place them in off-site storage.

You can vary the procedure for creating an archival backup by creating a stored script or shell script that updates dynamically. When you run the script, you can dynamically set the name of the restore point, backup format, and so on.

See Also:

Making an Archival Backup

This scenario makes a long-term archival backup with a backup tag of QUARTERLY and assigns it to a special family of Oracle Secure Backup tapes reserved for long-term storage. Note the following features of this example:

  • The FOREVER keyword indicates that this backup is never eligible for deletion by the backup retention policy.

  • The BACKUP command creates the restore point named FY06Q4 to match the SCN at which point this backup is consistent.

To make a long-term archival backup:

  1. Start RMAN and connect to a target database and recovery catalog.

    The target database can be open or mounted. A recovery catalog is required for KEEP FOREVER, but is not required for any other KEEP option.

  2. Run BACKUP ... KEEP to make the backup.

    The following example generates a data file and archived log backup and creates a normal restore point. The specified restore point must not already exist.

    The log backup contains just those archived logs needed to restore this backup to a consistent state. The database performs an online redo log switch to archive the redo that is in the current online logs and is necessary to make this new backup consistent. The control file autobackup has a copy of the restore point, so it can be referenced as soon as the control file is restored.

    RUN
    {
      ALLOCATE CHANNEL c1 
        DEVICE TYPE sbt
        PARMS 'ENV=(OB_MEDIA_FAMILY=archival_backup)';
      BACKUP DATABASE
        TAG quarterly
        KEEP FOREVER
        RESTORE POINT FY06Q4;
    }
    

    The following variation keeps the backup for 365 days instead of keeping it forever. After a year has passed, the backup becomes obsolete regardless of the backup retention policy settings.

    RUN
    {
      ALLOCATE CHANNEL c1 DEVICE TYPE sbt
        PARMS 'ENV=(OB_MEDIA_FAMILY=archival_backup)';
      BACKUP DATABASE
        TAG quarterly
        KEEP UNTIL TIME 'SYSDATE+365'
        RESTORE POINT FY06Q4;
    }
    

Making a Temporary Archival Backup

One purpose of an archival backup is to create a test database. The technique for making a test database is essentially the same as the technique described in "Making an Archival Backup for Long-Term Storage". The difference is that you intend to delete the backup soon after creating it.

You can specify the temporary status of the backup with the BACKUP ... KEEP UNTIL parameter. Assume that you want to make a backup and then restore it to a new host the same day. In this case, you can specify KEEP UNTIL TIME SYSDATE+1 to indicate that RMAN overrides the retention policy for this backup for only one day. After one day, the backup becomes obsolete, regardless of any configured backup retention policy.

The command in Example 9-11 makes an archival backup on a temporary disk with the tag TESTDB. The example creates a normal restore point, which is a label for the time to which the backup is recovered. RMAN only backs up the archived redo logs if the database is open during the backup. Archived logs are not needed for offline backups and so are not backed up.

Example 9-11 Creating a Temporary Archival Backup

BACKUP DATABASE
  FORMAT '/disk1/oraclebck/%U' 
  TAG TESTDB 
  KEEP UNTIL TIME 'SYSDATE+1' 
  RESTORE POINT TESTDB06;

The recommended technique for restoring an archival backup is to use the DUPLICATE command. See "Backup-Based Duplication Without a Target Connection: Example".

Backing Up RMAN Backups

This section explains how to back up backup sets and image copies.

About Backups of Backups

You can use the BACKUP BACKUPSET command to back up backup sets produced by other backup jobs. You can also use BACKUP RECOVERY AREA to back up recovery files created in the current and all previous fast recovery area destinations. Recovery files are full and incremental backup sets, control file autobackups, data file copies, and archived redo logs. SBT and disk backups are supported for BACKUP RECOVERY AREA. For disk backups of the recovery files, you must use the TO DESTINATION option.

The preceding commands are especially useful in the following scenarios:

  • Ensuring that all backups exist both on disk and on tape.

  • Moving backups from disk to tape and then freeing space on disk. This task is especially important when the database uses a fast recovery area so that the space can be reused as needed.

You can also use the BACKUP COPY OF command to back up image copies of data files, control files, and archived redo logs. The output of this command can be either backup sets or image copies, so you can generate backup sets from image copies. This form of backup is used to back up a database backup created as image copies on disk to tape.

Multiple Copies of Backup Sets

The BACKUP BACKUPSET command creates additional copies of backup pieces in a backup set, but does not create a new backup set. Thus, BACKUP BACKUPSET is similar to using the DUPLEX or MAXCOPIES option of BACKUP (see "Duplexing Backup Sets"). The extra copy of a backup set created by BACKUP BACKUPSET is not a new backup set, just as copies of a backup set produced by other forms of the BACKUP command are not separate backup sets.

Viewing the Effect of a Backup Retention Policy on Backups of Backups

For a backup retention policy based on redundancy, a backup set is counted as one instance of a backup. This statement is true even if there are multiple copies of the backup pieces that form the backup set, such as when a backup set has been backed up from disk to tape.

For a recovery window retention policy, either all of the copies of a backup set are obsolete, or none of them are. This point is easiest to grasp when viewing the output of the LIST and REPORT commands.

To view the effect of a backup retention policy on backups of backups:

  1. Back up a data file.

    The following example backs up data file 5:

    BACKUP AS BACKUPSET DATAFILE 5;
    
  2. Run the LIST command for the data file backup from Step 1.

    For example, run the following command (sample output included).

    
    LIST BACKUP OF DATAFILE 5 SUMMARY;
     
    List of Backups
    ===============
    Key     TY LV S Device Type Completion Time #Pieces #Copies Compressed Tag
    ------- -- -- - ----------- --------------- ------- ------- ---------- ---
    18      B  F  A DISK        04-AUG-13       1       1       NO         TAG20070804T160 134
    
  3. Use the backup set key from the previous step to back up the backup set.

    For example, enter the following command:

    BACKUP BACKUPSET 18;
    
  4. Run the same LIST command that you ran in Step 2.

    For example, run the following command (sample output included).

    
    LIST BACKUP OF DATAFILE 5 SUMMARY;
     
    List of Backups
    ===============
    Key     TY LV S Device Type Completion Time #Pieces #Copies Compressed Tag
    ------- -- -- - ----------- --------------- ------- ------- ---------- ---
    18      B  F  A DISK        04-AUG-13       1       2       NO         TAG20070804T160 134
    

    Only one backup set is shown in this output, but there are now two copies of it.

  5. Generate a report to see the effect of these copies under a redundancy-based backup retention policy.

    For example, issue the following command:

    REPORT OBSOLETE REDUNDANCY 1;
     
    

    No copy is reported as obsolete because both copies of the backup set have the same values for set_stamp and set_count.

  6. Generate a report to see the effect of these copies under a recovery window-based backup retention policy.

    For example, issue the following command:

    REPORT OBSOLETE RECOVERY WINDOW OF 1 DAYS;
     
    

    No copy of the backup set is reported as obsolete or based on the CHECKPOINT_CHANGE# of this backup set, with the current time and the availability of other backups.

    See Also:

Backing Up Backup Sets with RMAN

This section explains how to use the BACKUP BACKUPSET command to copy backup sets from disk to tape. The procedure assumes that you have configured an SBT device as your default device.

To back up backup sets from disk to tape:

  1. If you are backing up a subset of available backup sets, then execute the LIST BACKUPSET command to obtain their primary keys.

    The following example lists the backup sets in summary form:

    RMAN> LIST BACKUPSET SUMMARY;
    
    List of Backups
    ===============
    Key TY LV S Device Type Completion Time #Pieces #Copies Comp Tag
    --- -- -- - ----------- --------------- ------- ------- ---- ---
    1   B  F  A DISK        28-MAY-13       1       1       NO   TAG20070528T132432
    2   B  F  A DISK        29-MAY-13       1       1       NO   TAG20070529T132433
    3   B  F  A DISK        30-MAY-13       1       1       NO   TAG20070530T132434
    

    The following example lists details about backup set 3:

    RMAN> LIST BACKUPSET 3;
    
    List of Backup Sets
    ===================
    
    BS Key  Type LV Size       Device Type Elapsed Time Completion Time
    ------- ---- -- ---------- ----------- ------------ ---------------
    3       Full    8.33M      DISK        00:00:01     30-MAY-13
            BP Key: 3   Status: AVAILABLE  Compressed: NO  Tag: TAG20070530T132434
            Piece Name: /disk1/oracle/dbs/c-35764265-20070530-02
      Control File Included: Ckp SCN: 397221       Ckp time: 30-MAY-13
      SPFILE Included: Modification time: 30-MAY-13
      SPFILE db_unique_name: PROD
    
  2. Execute the BACKUP BACKUPSET command.

    The following example backs up all disk backup sets to tape and then deletes the input disk backups:

    BACKUP BACKUPSET ALL
      DELETE INPUT;
    

    The following example backs up only the backup sets with the primary key 1 and 2 to tape and then deletes the input disk backups:

    BACKUP BACKUPSET 1,2
      DELETE INPUT;
    
  3. Optionally, execute the LIST command to see a listing of backup sets and pieces.

    The output lists all copies, including backup piece copies created by BACKUP BACKUPSET.

Backing Up Image Copy Backups with RMAN

This section explains how to use the BACKUP command to back up image copies to tape. It is assumed that you have configured an SBT device as your default device.

When you back up image copies that have multiple copies of the data files, specifying tags for the backups makes it easier to identify the input image copy. All image copies of data files have tags. The tag of an image copy is inherited by default when the image copy is backed up as a new image copy.

To back up image copies from disk to tape:

  1. Issue the BACKUP ... COPY OF or BACKUP DATAFILECOPY command.

    The following example backs up data file copies that have the tag DBCopy:

    BACKUP DATAFILECOPY FROM TAG monDBCopy;
    

    The following example backs up the latest image copies of a database to tape, assigns the tag QUARTERLY_BACKUP, and deletes the input disk backups:

    BACKUP DEVICE TYPE sbt
      TAG "quarterly_backup"
      COPY OF DATABASE 
      DELETE INPUT;
    
  2. Optionally, issue the LIST command to see a listing of backup sets. The output lists all copies, including backup piece copies created by the BACKUP command with the BACKUPSET clause.