|Oracle9i Recovery Manager User's Guide
Release 2 (9.2)
Part Number A96566-01
This chapter describes how to get started using RMAN. This chapter contains these topics:
?/rdbms/demo subdirectory (the location differs depending on your operating system) contains a number of sample RMAN scripts. The scripts are suffixed with the
.rcv filename extension.
These files are case studies of RMAN commands that are fully documented so that you can understand the features used. You can customize them for your use in your own backup and recovery plan, or simply read them to learn how RMAN is used.
As explained in "Authentication for Database Connections", to run RMAN commands you must first connect RMAN to the target database with the
SYSDBA role to a dedicated server process. You can either connect using operating system authentication or using Oracle Net. An Oracle Net connection as
SYSDBA requires that you create a password file.
|If you connect using . . .||Then . . .|
Operating system authentication
See Also: Oracle9i Database Administrator's Guide to learn about authentication, and your operating system specific documentation to learn about adding
After you have learned how to connect to a target database, you can immediately begin performing backup and recovery operations. Use the examples in this section to go through a basic backup and restore scenario using a test database. These examples assume the following:
This section contains these topics:
The first task is to connect to the target database. If you have created a recovery catalog, then you can connect to it as well--although these examples assume you are connecting in the default
At the operating system command line, enter the following to connect to the target database in the default
If the database is already mounted or open, then RMAN displays output similar to the following:
DBID value displayed is the database identifier for the target database.
If the target database is not started, then RMAN shows the following message:
Oracle9i Recovery Manager Reference for connection options
If the database is not started, then run the
startup command at the RMAN prompt, specifying an initialization parameter file only if you do not use a server parameter file. This example starts the instance with the server parameter file:
If the database is open, then you can run the following RMAN commands to close it cleanly and then mount it:
RMAN> SHUTDOWN IMMEDIATE database closed database dismounted Oracle instance shut down RMAN> STARTUP MOUNT
In this example, you generate a report describing the target datafiles. Run the
report schema command as follows:
RMAN displays the datafiles currently in the target database. Depending on the contents of the database, you will see output similar to the following:
Report of database schema File K-bytes Tablespace RB segs Datafile Name ---- ---------- -------------------- ------- ------------------- 1 204800 SYSTEM *** /oracle/oradata/trgt/system01.dbf 2 20480 UNDOTBS *** /oracle/oradata/trgt/undotbs01.dbf 3 10240 CWMLITE *** /oracle/oradata/trgt/cwmlite01.dbf 4 10240 DRSYS *** /oracle/oradata/trgt/drsys01.dbf 5 10240 EXAMPLE *** /oracle/oradata/trgt/example01.dbf 6 10240 INDX *** /oracle/oradata/trgt/indx01.dbf 7 10240 TOOLS *** /oracle/oradata/trgt/tools01.dbf 8 10240 USERS *** /oracle/oradata/trgt/users01.dbf
In this task, you back up the database to the default disk location. The default location is port-specific. For example, on most UNIX systems the location is
?/dbs. Because you do not specify the
format parameter in this example, RMAN assigns the backup a unique filename. If you do not manually allocate a channel, then RMAN uses a preconfigured disk channel by default.
You can make two basic types of backups: full and incremental. In a full backup, RMAN backs up all blocks of the target database files. In an incremental backup, RMAN backs up only the blocks that have changed since a previous backup.
backup command at the RMAN prompt as follows to make a full backup of the datafiles, control file, and current server parameter file (if the instance is started with a server parameter file) to the default device type:
Unless you explicitly configure the default device to tape with the
CONFIGURE command, the default device is disk.
When you run the
BACKUP command, RMAN creates a backup set, which is a logical object that contains one or more backup pieces. The
backup command output contains the essential information about the backup, as shown in the following example:
Starting backup at OCT 12 2001 19:09:48 using target database controlfile instead of recovery catalog allocated channel: ORA_DISK_1 channel ORA_DISK_1: sid=10 devtype=DISK channel ORA_DISK_1: starting full datafile backupset channel ORA_DISK_1: specifying datafile(s) in backupset including current SPFILE in backupset including current controlfile in backupset input datafile fno=00001 name=/oracle/oradata/trgt/system01.dbf input datafile fno=00002 name=/oracle/oradata/trgt/undotbs01.dbf input datafile fno=00003 name=/oracle/oradata/trgt/cwmlite01.dbf input datafile fno=00004 name=/oracle/oradata/trgt/drsys01.dbf input datafile fno=00005 name=/oracle/oradata/trgt/example01.dbf input datafile fno=00006 name=/oracle/oradata/trgt/indx01.dbf input datafile fno=00007 name=/oracle/oradata/trgt/tools01.dbf input datafile fno=00008 name=/oracle/oradata/trgt/users01.dbf channel ORA_DISK_1: starting piece 1 at OCT 12 2001 19:09:56 channel ORA_DISK_1: finished piece 1 at OCT 12 2001 19:10:31 piece handle=/oracle/dbs/lvd6dtk1_1_1 comment=NONE channel ORA_DISK_1: backup set complete, elapsed time: 00:00:39 Finished backup at OCT 12 2001 19:10:33
Incremental backups are a convenient way to conserve storage space because they back up only database blocks that have changed. RMAN compares the current datafiles to a base backup, also called a level 0 backup, to determine which blocks to back up.
For example, you can make a full backup as a base backup and then make some updates to the test database and commit them. Then, when you run the following command, RMAN backs up only those blocks that have changed since the previous full backup:
Note that you will see lines such as the following in the output:
This line does not indicate a problem, but simply means that no base
0 incremental backup exists. RMAN automatically creates a
0 backup for its base incremental backup.
Besides backing up the whole database, you can back up a tablespace or datafile. In this example, back up the
SYSTEM tablespace to disk using the preconfigured disk channel. Of course, you can choose to back up a different object.
backup command at the RMAN prompt as follows:
"Backing Up Tablespaces with RMAN" to learn how to back up tablespaces
Typically, database administrators back up archived logs on disk to a third-party storage medium such as tape. You can also back up archived logs to disk. In either case, you can delete the input logs automatically after the backup completes.
To back up all archived logs and delete the input logs (from the primary archiving destination only), run the
backup command at the RMAN prompt as follows:
"Backing Up Archived Redo Logs with RMAN" to learn how to back up archived redo logs
In this example, make an image copy of datafile 1 to a new location. An image copy differs from a backup set in that it is not in an RMAN-specific format. It is the equivalent of a copy made using an operating system command such as the UNIX
This example uses an automatically allocated disk channel to create a datafile copy named
df1.bak. Run the
copy command as follows from the RMAN prompt, specifying the path name for the backup:
RMAN displays the full filename of the created file in the output, as in this example:
Starting copy at OCT 12 2001 19:11:28 using channel ORA_DISK_1 channel ORA_DISK_1: copied datafile 1 output filename=/tmp/df1.cpy recid=141 stamp=442955509 Finished copy at OCT 12 2001 19:11:59 Finished copy at 18-APR-01
To list the backup sets and image copies that you have created, run the
list command as follows:
RMAN displays which backup sets and pieces it created as well as which datafiles it included in those sets, as in the following example:
BS Key Type LV Size Device Type Elapsed Time Completion Time ------- ---- -- ---------- ----------- ------------ -------------------- 699 Full 122M DISK 00:00:30 OCT 12 2001 19:10:23 BP Key: 1293 Status: AVAILABLE Tag: TAG20011012T191001 Piece Name: /oracle/dbs/lvd6dtk1_1_1 Controlfile Included: Ckp SCN: 322170 Ckp time: OCT 12 2001 19:09:53 SPFILE Included: Modification time: OCT 12 2001 19:09:53 List of Datafiles in backup set 699 File LV Type Ckp SCN Ckp Time Name ---- -- ---- ---------- -------------------- ---- 1 Full 322172 OCT 12 2001 19:09:56 /oracle/oradata/trgt/system01.dbf 2 Full 322172 OCT 12 2001 19:09:56 /oracle/oradata/trgt/undotbs01.dbf 3 Full 322172 OCT 12 2001 19:09:56 /oracle/oradata/trgt/cwmlite01.dbf 4 Full 322172 OCT 12 2001 19:09:56 /oracle/oradata/trgt/drsys01.dbf 5 Full 322172 OCT 12 2001 19:09:56 /oracle/oradata/trgt/example01.dbf 6 Full 322172 OCT 12 2001 19:09:56 /oracle/oradata/trgt/indx01.dbf 7 Full 322172 OCT 12 2001 19:09:56 /oracle/oradata/trgt/tools01.dbf 8 Full 322172 OCT 12 2001 19:09:56 /oracle/oradata/trgt/users01.dbf
To list image copies, run the following command:
RMAN displays both datafile and control file copies as well as archived redo logs (an archived redo log is considered a type of copy):
List of Datafile Copies Key File S Completion Time Ckp SCN Ckp Time Name ------- ---- - -------------------- ---------- -------------------- ---- 141 1 A OCT 12 2001 19:11:49 322204 OCT 12 2001 19:11:29 /tmp/df1.copy List of Archived Log Copies Key Thrd Seq S Low Time Name ------- ---- ------- - -------------------- ---- 1105 1 70 A OCT 10 2001 17:55:49 /oracle/oradata/trgt/arch/archive1_70.dbf 1106 1 71 A OCT 11 2001 05:12:06 /oracle/oradata/trgt/arch/archive1_71.dbf 1107 1 72 A OCT 11 2001 14:54:36 /oracle/oradata/trgt/arch/archive1_72.dbf 1108 1 73 A OCT 11 2001 14:55:48 /oracle/oradata/trgt/arch/archive1_73.dbf 1109 1 74 A OCT 11 2001 15:13:27 /oracle/oradata/trgt/arch/archive1_74.dbf 1110 1 75 A OCT 12 2001 06:56:22 /oracle/oradata/trgt/arch/archive1_75.dbf
Check that you are able to restore the backups that you created without actually restoring them. Run the
VALIDATE command as follows:
You should see output similar to the following:
Starting restore at 07-DEC-01 allocated channel: ORA_DISK_1 channel ORA_DISK_1: sid=9 devtype=DISK channel ORA_DISK_1: starting validation of datafile backupset channel ORA_DISK_1: restored backup piece 1 piece handle=/oracle/dbs/07db39tr_1_1 tag=TAG20011012T191001 params=NULL channel ORA_DISK_1: validation complete Finished restore at 07-DEC-01
If there are no error messages, then RMAN confirms that the database backup can be restored. When there is an error, RMAN always displays an error banner and provides messages indicating the nature of the error. For example, if you attempt to restore the database when some backups do not exist, then RMAN displays an error stack such as the following:
RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03002: failure of restore command at 12/07/2001 17:28:24 RMAN-06026: some targets not found - aborting restore RMAN-06023: no backup or copy of datafile 8 found to restore RMAN-06023: no backup or copy of datafile 7 found to restore
The primary aspect of developing a backup and recovery strategy is learning what to do in case of a media failure. In this scenario, you simulate a media failure. First, shut down the database and then exit RMAN with the following commands:
After the database is shut down, use the output from the
SCHEMA command in "Reporting the Current Schema of the Target Database" to identify the filenames of your datafiles. Temporarily rename some or all of your database files with operating system commands (but make sure not to rename your control files). This action simulates a media failure because Oracle is able to find the datafiles during startup.
This UNIX example temporarily renames the datafile in the
Now, start RMAN and attempt to open the database as in the following example:
% rman TARGET / RMAN> STARTUP RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03002: failure of startup command at 08/29/2001 13:46:04 ORA-01157: cannot identify/lock data file 7 - see DBWR trace file ORA-01110: data file 7: '/oracle/oradata/trgt/tools01.dbf'
The database mounts but does not open because some datafiles require recovery. To restore and recover the database using the default disk channel, simply run the following commands from the RMAN prompt:
RMAN uses the backups and copies that you made earlier and restores the files to their default locations. Then, it uses archived redo logs (if needed) to recover the database. After recovery is complete, open the database:
It is not uncommon for a media failure to affect some but not all files in a database. In this scenario, you simulate a media failure to a datafile in a single tablespace. Use the output from the
SCHEMA command in the previous example to identify the filenames of datafiles in a tablespace other than the
SYSTEM tablespace. This example renames datafiles in the
While the database is open, rename one or more datafiles in the tablespace that you selected. This UNIX example temporarily renames a datafile in
If you attempt to update a table located in one of the renamed datafiles, Oracle generates an error message because the datafile is unavailable. You need to restore and recover the missing datafiles. First, start RMAN and then take the tablespace offline using the
SQL command, making sure to specify the name of the tablespace:
To restore and recover the tablespace with the default disk channel, simply run the following commands from the RMAN prompt:
After recovery is complete, bring the tablespace online with the following command:
RMAN contains some default configuration settings. These settings apply to all RMAN sessions until you explicitly change or disable them with another
CONFIGURE command. You can always return to a default configuration setting by running the
Of the possible configurations, perhaps the most important are the automatic channels and the retention policy. RMAN is preconfigured with an automatic disk channel so that you can make backups and copies to disk without manually allocating channels.
To see all the current RMAN configuration settings, run this command:
Chapter 8, "Configuring the Recovery Manager Environment" to learn how to create RMAN configurations
By default, RMAN connects to the target database in
NOCATALOG mode, meaning that it uses the control file in the target database as the sole repository of RMAN metadata. Perhaps the most important decision you make when using RMAN is whether to create a recovery catalog as the RMAN repository for normal production operations. A recovery catalog is a schema created in a separate database that contains metadata obtained from the target control file.
In general, Oracle Corporation advises using a catalog when you manage multiple databases. If you have more than one database to back up, then you can create one systemwide recovery catalog and store metadata for all the databases in this catalog. Hence, you avoid the extra space requirements and memory overhead of maintaining multiple databases, each with a single catalog. You need to take extra precautions when backing up the catalog, however, because if you lose the catalog then you lose the metadata for multiple target databases.
This section outlines some of the costs and benefits associated with using and not using a recovery catalog. If you decide to create a catalog, refer to "Creating the Recovery Catalog" for instructions.
"RMAN Repository" for an overview of the function of the RMAN repository in the RMAN environment
When you use a recovery catalog, RMAN can perform a wider variety of automated backup and recovery functions than when you use the control file in the target database as the sole repository of metadata. The following features are available only with a catalog:
CONTROL_FILE_RECORD_KEEP_TIMEsetting, you can keep historical metadata.
The main cost of using a catalog is the maintenance overhead required for this additional database. For example, you have to:
Hence, unless you manage a network of databases, you may choose to avoid the overhead and use the control file as the exclusive repository of metadata. When you use a control file as the RMAN repository, RMAN still functions effectively. If you do not use a catalog, read the section "Managing the RMAN Repository Without a Recovery Catalog". Specifically, make sure you:
Chapter 16, "Managing the Recovery Manager Repository" to learn how to manage the recovery catalog