20 Performing RMAN Recovery: Advanced Scenarios

If you have configured control file autobackups, then the server parameter file is backed up with the control file whenever an autobackup is taken.

To restore the server parameter file from the control file autobackup, you must first set the DBID for your database and then use the RESTORE SPFILE FROM AUTOBACKUP command. If the autobackup is in a nondefault format, then first use the SET CONTROLFILE AUTOBACKUP FORMAT command to specify the format.

Example 20-1 sets the DBID and restores the server parameter file from a control file autobackup in a nondefault location.

You can also restore the server parameter file as a client-side initialization parameter file with the TO PFILE 'filename' clause. The file name that you specify must be on a file system accessible from the host where the RMAN client is running. This file need not be accessible directly from the host running the instance.

The following RMAN command creates an initialization parameter file named /tmp/initTEMP.ora on the system running the RMAN client:

RESTORE SPFILE TO PFILE '/tmp/initTEMP.ora';

To restart the instance with the initialization parameter file, use the following command, again running RMAN on the same client host:

STARTUP FORCE PFILE='/tmp/initTEMP.ora';

This section assumes that you have RMAN backups of the control file, but do not use a recovery catalog. It also assumes that you enabled the control file autobackup feature for the target database and can restore an autobackup of the control file.

Because the autobackup uses a well-known format, RMAN can restore it even though it does not have a repository available that lists the available backups. You can restore the autobackup to the default or a new location. RMAN replicates the control file to all CONTROL_FILES locations automatically.

Because you are not connected to a recovery catalog, the RMAN repository contains only information about available backups at the time of the control file backup. If you know the location of other usable backup sets or image copies, then add them to the control file RMAN repository with the CATALOG command.

To recover the database with a control file autobackup in NOCATALOG mode:

1. Start RMAN and connect to a target database.

   See Also:

   "Making Database Connections with RMAN"

2. Start the target database instance without mounting the database. For example:

   STARTUP NOMOUNT;
   

3. Set the database identifier for the target database with the SET DBID command.

   RMAN displays the DBID whenever you connect to a target database. You can also obtain it by inspecting saved RMAN log files, querying the catalog, or looking at the file names of control file autobackup. For example, run:

   SET DBID 676549873;
   

4. Write an RMAN command file to restore the autobackup control file and perform recovery.

   The command file contains the following steps:

   1. Optionally, specify the most recent backup time stamp that RMAN can use when searching for a control file autobackup to restore.

   2. If you know that a different control file autobackup format was in effect when the control file autobackup was created, then specify a nondefault format for the restore of the control file.

   3. If an SBT channel created the control file autobackup, then allocate one or more SBT channels. Because no recovery catalog is available, you cannot use preconfigured channels.

   4. Restore the autobackup of the control file, optionally setting the maximum number of days backward that RMAN can search and the initial sequence number that it uses in its search for the first day.

   5. If you know that the control file contains information about configured channels that is useful to you in the rest of the restore process, then you can exit RMAN to clear manually allocated channels from Step 4.c.

      If you restart the RMAN client and mount the database, then these configured channels are available for your use. If you do not care about using configured channels from your control file, then you can simply mount the database.

   6. This step depends on whether the online redo logs are available. The option OPEN RESETLOGS is always required after recovery with a backup control file, regardless of whether logs are available.

      If the online redo logs are usable, then RMAN can find and apply these logs. Perform a complete restore and recovery as described in "Performing Complete Database Recovery".

      If the online redo logs are unusable, then perform DBPITR as described in "Performing Database Point-in-Time Recovery". An UNTIL clause is required to specify a target time, SCN, or log sequence number for the recovery before the first SCN of the online redo logs (otherwise, RMAN issues the RMAN-6054 error).

      When you perform DBPITR with a backup control file, before opening the database with RESETLOGS, you can open the database read-only using SQL*Plus and run queries as needed to verify that the effects of the logical corruption have been reversed. If you are satisfied with the results, then you can open the database with RESETLOGS.

      Note:

      When specifying log sequences, if the last created archived redo log has sequence n, then specify UNTIL SEQUENCE n+1 so that RMAN applies n and then stops.

   In the following example, the online redo log files have been lost, and the most recent archived redo log sequence number is 13243. This example shows how to restore the control file autobackup and recover through the latest log.

   RUN 
   {
     # Optionally, set upper limit for eligible time stamps of control file 
     # backups
     # SET UNTIL TIME '09/10/2013 13:45:00';
     # Specify a nondefault autobackup format only if required
     # SET CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK 
     #   TO '?/oradata/%F.bck';
     ALLOCATE CHANNEL c1 DEVICE TYPE sbt PARMS '...'; # allocate manually
     RESTORE CONTROLFILE FROM AUTOBACKUP
       MAXSEQ 100           # start at sequence 100 and count down
       MAXDAYS 180;         # start at UNTIL TIME and search back 6 months
     ALTER DATABASE MOUNT;
   }
   # Now use automatic channels configured in restored control file
   RESTORE DATABASE UNTIL SEQUENCE 13244;
   RECOVER DATABASE UNTIL SEQUENCE 13244;
   

5. If recovery was successful, then open the database and reset the online logs:

   ALTER DATABASE OPEN RESETLOGS;
   

The procedure for disaster recovery is similar to the procedure for recovering the database with a backup control file in NOCATALOG mode. If you are restoring the database to a new host, then review the considerations described in "Restoring a Database on a New Host".

This scenario assumes that the Linux server on which your database was running has been damaged beyond repair. Fortunately, you backed up the database to Oracle Secure Backup and have the tapes available. The scenario assumes the following:

• Oracle Database is installed on the new host.

• You are restoring the database to a new Linux host with the same directory structure as the old host.

• You have one tape drive containing backups of all the data files and archived redo logs through log 1124, and autobackups of the control file and server parameter file.

• You do not use a recovery catalog with the database.

To recover the database on the new host:

1. Ensure that the prerequisites described in Prerequisites of Disaster Recovery are met.

2. If possible, restore or re-create all relevant network files such as tnsnames.ora and listener.ora and a password file.

3. Start RMAN and connect to the target database instance.

   At this stage, no initialization parameter file exists. If you have set ORACLE_SID and ORACLE_HOME, then you can use operating system authentication to connect as SYSDBA or SYSBACKUP.

   See Also:

   "Making Database Connections with RMAN"

4. Specify the DBID for the target database with the SET DBID command, as described in "Restoring the Server Parameter File".

   For example, enter the following command:

   SET DBID 676549873;
   

5. Run the STARTUP NOMOUNT command.

   When the server parameter file is not available, RMAN attempts to start the instance with a dummy server parameter file.

6. Allocate a channel to the media manager and then restore the server parameter file from autobackup.

   For example, enter the following command to restore the server parameter file from Oracle Secure Backup:

   RUN
   {
     ALLOCATE CHANNEL c1 DEVICE TYPE sbt;
     RESTORE SPFILE FROM AUTOBACKUP;
   }
   

7. Restart the instance with the restored server parameter file.

   STARTUP FORCE NOMOUNT;
   

8. Write a command file to perform the restore and recovery operation, and then execute the command file. The command file must do the following:

   1. Allocate a channel to the media manager.

   2. Restore a control file autobackup (see "Performing Recovery with a Backup Control File and No Recovery Catalog").

   3. Mount the restored control file.

   4. Catalog any backups not recorded in the repository with the CATALOG command.

   5. Restore the data files to their original locations. If volume names have changed, then run SET NEWNAME commands before the restore operation and perform a switch after the restore operation to update the control file with the new locations for the data files, as shown in the following example.

   6. Recover the data files. RMAN stops recovery when it reaches the log sequence number specified.

   RMAN> RUN
   {
     # Manually allocate a channel to the media manager
     ALLOCATE CHANNEL t1 DEVICE TYPE sbt;
     # Restore autobackup of the control file. This example assumes that you have 
     # accepted the default format for the autobackup name.
     RESTORE CONTROLFILE FROM AUTOBACKUP;
     #  The set until command is used in case the database
     #  structure has changed in the most recent backups, and you want to
     #  recover to that point in time. In this way RMAN restores the database
     #  to the same structure that the database had at the specified time.
     ALTER DATABASE MOUNT;
     SET UNTIL SEQUENCE 1124 THREAD 1;
     RESTORE DATABASE;
     RECOVER DATABASE;
   }
   

   The following example of the RUN command shows the same scenario except with new file names for the restored data files:

   RMAN> RUN
   {
     #  If you must restore the files to new locations,
     #  use SET NEWNAME commands:
     SET NEWNAME FOR DATAFILE 1 TO '/dev/vgd_1_0/rlvt5_500M_1';
     SET NEWNAME FOR DATAFILE 2 TO '/dev/vgd_1_0/rlvt5_500M_2';
     SET NEWNAME FOR DATAFILE 3 TO '/dev/vgd_1_0/rlvt5_500M_3';
     ALLOCATE CHANNEL t1 DEVICE TYPE sbt;
     RESTORE CONTROLFILE FROM AUTOBACKUP;
     ALTER DATABASE MOUNT;
     SET UNTIL SEQUENCE 124 THREAD 1;
     RESTORE DATABASE;
     SWITCH DATAFILE ALL; # Update control file with new location of data files.
     RECOVER DATABASE;
   }
   

9. If recovery was successful, then open the database and reset the online logs:

   ALTER DATABASE OPEN RESETLOGS;
   

To prepare for the restoration of the database to a new host, take the following steps:

• Record the DBID for your source database. If you do not know the DBID for your database, then see "Determining the DBID of the Database" to learn how to determine the DBID.

• Make the source database initialization parameter file accessible on the new host. Copy the file from the old host to a new host by using an operating system utility.

• If you perform a test restore operation only, then ensure that RMAN is not connected to the recovery catalog. Otherwise, RMAN records metadata about the restored data files in the recovery catalog. This metadata interferes with future attempts to restore and recover the primary database.

  If you must use a recovery catalog because the control file is not large enough to contain the RMAN repository data on all of the backups that you must restore, then use Oracle Data Pump to export the catalog and import it into a different schema or database. Afterward, use the copied recovery catalog for the test restore. Otherwise, the recovery catalog considers the restored database as the current target database.

• Ensure that backups used for the restore operation are accessible on the restore host. For example, if the backups were made with a media manager, then verify that the tape device is connected to the new host. If you are using disk copies, then use the procedure in the following section.

• If you are performing a trial restore of the production database, then perform either of the following actions before restoring the database in the test environment:

  • If the test database will use a fast recovery area that is physically different from the recovery area used by the production database, then set DB_RECOVERY_FILE_DEST in the test database instance to the new location.

  • If the test database will use a fast recovery area that is physically the same as the recovery area used by the production database, then set DB_UNIQUE_NAME in the test database instance to a different name from the production database.

  If you do not perform either of the preceding actions, then RMAN assumes that you are restoring the production database and deletes flashback logs from the fast recovery area because they are considered unusable.

To move the database to a new host by using data file copies or backup sets on disk, you must transfer the files manually to the new host. This example assumes that RMAN is using a recovery catalog.

To restore backup files to a new host:

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

   See Also:

   "Making Database Connections with RMAN"

2. Run a LIST command to see a listing of backups of the data file and control file autobackups.

   For example, enter the following command to view data file copies:

   LIST COPY;
   

   For example, enter the following command to view control file backups:

   LIST BACKUP OF CONTROLFILE;
   

   The piece name of the autobackup must use the %F substitution variable, so the autobackup piece name includes the string c-IIIIIIIIII-YYYYMMDD-QQ, where IIIIIIIIII stands for the DBID, YYYYMMDD is a time stamp in the Gregorian calendar of the day the backup is generated, and QQ is the sequence in hexadecimal.

3. Copy the backups to the new host with an operating system utility.

   Enter a command such as the following to copy all data file copies to the ?/oradata/trgt directory on the new host:

   % cp -r /disk1/*dbf /net/new_host/oracle/oradata/trgt
   

   Enter a command such as the following to copy the autobackup backup piece to the /tmp directory on the new host:

   % cp -r /disk1/auto_bkp_loc/c-1618370911-20130208-00 /net/new_host/tmp
   

   As explained in "Restoring the Server Parameter File from a Control File Autobackup", you must use the SET CONTROLFILE AUTOBACKUP FORMAT command when restoring an autobackup from a nondefault location.

This scenario tests whether you can restore your database to a new host. In this scenario, you have two networked Linux hosts, hosta and hostb. A target database named trgta is on hosta and is registered in recovery catalog catdb. You want to test the restore and recovery of trgta on hostb, while keeping database trgta up and running on hosta.

Now, assume that the directory structure of hostb is different from hosta. The target database is located in /net/hosta/dev3/oracle/dbs, but you want to restore the database to /net/hostb/oracle/oradata/test. You have tape backups of data files, control files, archived redo logs, and the server parameter file on a media manager accessible by both hosts. The ORACLE_SID for the TRGTA database is TRGTA and does not change for the restored database.

To restore the database on a new host:

1. Ensure that the backups of the target database are accessible on the new host.

   To test disaster recovery, you must have a recoverable backup of the target database. When preparing your disaster recovery strategy, ensure that the backups of the data files, control files, and server parameter file are restorable on hostb. Thus, you must configure the media management software so that hostb is a media manager client and can read the backup sets created on hosta. Consult the media management vendor for support on this issue.

2. Configure the ORACLE_SID on hostb.

   This scenario assumes that you want to start the RMAN client on hostb and authenticate yourself through the operating system. However, you must be connected to hostb either locally or through a net service name.

   After logging in to hostb with administrator privileges, edit the /etc/group file so that you are included in the DBA group:

   dba:*:614:<your_user_name>
   

   Set the ORACLE_SID environment variable on hostb to the same value used on hosta:

   % setenv ORACLE_SID trgta
   

3. Start RMAN on hostb and connect to the target database without connecting to the recovery catalog.

   For example, enter the following command:

   % rman NOCATALOG
   RMAN> CONNECT TARGET /
   

   See Also:

   "Making Database Connections with RMAN"

4. Set the DBID and start the database instance without mounting the database.

   For example, run SET DBID to set the DBID, then run STARTUP NOMOUNT:

   SET DBID 1340752057;
   STARTUP NOMOUNT
   

   RMAN fails to find the server parameter file, which has not yet been restored, but starts the instance with a "dummy" file. Sample output follows:

   startup failed: ORA-01078: failure in processing system parameters
   LRM-00109: could not open parameter file '/net/hostb/oracle/dbs/inittrgta.ora'
   
   trying to start the Oracle instance without parameter files ...
   Oracle instance started
   

5. Restore and edit the server parameter file.

   Because you enabled the control file autobackup feature when making your backups, the server parameter file is included in the backup. If you are restoring an autobackup that has a nondefault format, then use the SET CONTROLFILE AUTOBACKUP FORMAT command to indicate the format.

   Allocate a channel to the media manager, then restore the server parameter file as a client-side parameter file and use the SET command to indicate the location of the autobackup (in this example, the autobackup is in /tmp):

   RUN
   {
     ALLOCATE CHANNEL c1 DEVICE TYPE sbt PARMS '...';
     SET CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO '/tmp/%F';
     RESTORE SPFILE 
       TO PFILE '?/oradata/test/inittrgta.ora' 
       FROM AUTOBACKUP;
     SHUTDOWN ABORT;
   }
   

6. Edit the restored initialization parameter file.

   Change any location-specific parameters, for example, those ending in _DEST, to reflect the new directory structure. For example, edit the following parameters:

     - IFILE
     - LOG_ARCHIVE_DEST_1
     - CONTROL_FILES
   

7. Restart the instance with the edited initialization parameter file.

   For example, enter the following command:

   STARTUP FORCE NOMOUNT PFILE='?/oradata/test/inittrgta.ora';
   

8. Restore the control file from an autobackup and then mount the database.

   For example, enter the following command:

   RUN 
   {
     ALLOCATE CHANNEL c1 DEVICE TYPE sbt PARMS '...';
     RESTORE CONTROLFILE FROM AUTOBACKUP;
     ALTER DATABASE MOUNT;
   }
   

   RMAN restores the control file to whatever locations you specified in the CONTROL_FILES initialization parameter.

9. Catalog the data file copies that you copied in "Restoring Disk Backups to a New Host", using their new file names or CATALOG START WITH (if you know all the files are in directories with a common prefix easily addressed with a CATALOG START WITH command). For example, run:

   CATALOG START WITH '/oracle/oradata/trgt/';
   

   If you want to specify files individually, then you can execute a CATALOG command as follows:

   CATALOG DATAFILECOPY
     '/oracle/oradata/trgt/system01.dbf', '/oracle/oradata/trgt/undotbs01.dbf', 
     '/oracle/oradata/trgt/cwmlite01.dbf', '/oracle/oradata/trgt/drsys01.dbf',
     '/oracle/oradata/trgt/example01.dbf', '/oracle/oradata/trgt/indx01.dbf', 
     '/oracle/oradata/trgt/tools01.dbf', '/oracle/oradata/trgt/users01.dbf';
   

10. Start a SQL*Plus session on the new database and query the database file names recorded in the control file.

    Because the control file is from the trgta database, the recorded file names use the original hosta file names. You can query V$ views to obtain this information. Run the following query in SQL*Plus:

    COLUMN NAME FORMAT a60
    SPOOL  LOG '/tmp/db_filenames.out'
    SELECT FILE# AS "File/Grp#", NAME 
    FROM   V$DATAFILE
    UNION
    SELECT GROUP#,MEMBER 
    FROM   V$LOGFILE;
    SPOOL OFF
    EXIT
    

11. Write the RMAN restore and recovery script. The script must include the following steps:

    1. For each data file on the destination host that is restored to a different path than it had on the source host, use a SET NEWNAME command to specify the new path on the destination host. If the file systems on the destination system are set up to have the same paths as the source host, then do not use SET NEWNAME for those files restored to the same path as on the source host.

    2. For each online redo log that is to be created at a different location than it had on the source host, use SQL ALTER DATABASE RENAME FILE commands to specify the path name on the destination host. If the file systems on the destination system are set up to have the same paths as the source host, then do not use ALTER DATABASE RENAME FILE for those files restored to the same path as on the source host.

    3. Perform a SET UNTIL operation to limit recovery to the end of the archived redo logs. The recovery stops with an error if no SET UNTIL command is specified.

    4. Restore and recover the database.

    5. Run the SWITCH DATAFILE ALL command so that the control file recognizes the new path names as the official new names of the data files.

    The following code shows the RMAN script reco_test.rman that can perform the restore and recovery operation.

    RUN
    {
      # allocate a channel to the tape device
      ALLOCATE CHANNEL c1 DEVICE TYPE sbt PARMS '...';
    
      # rename the data files and online redo logs
      SET NEWNAME FOR DATAFILE 1 TO '?/oradata/test/system01.dbf';
      SET NEWNAME FOR DATAFILE 2 TO '?/oradata/test/undotbs01.dbf';
      SET NEWNAME FOR DATAFILE 3 TO '?/oradata/test/cwmlite01.dbf';
      SET NEWNAME FOR DATAFILE 4 TO '?/oradata/test/drsys01.dbf';
      SET NEWNAME FOR DATAFILE 5 TO '?/oradata/test/example01.dbf';
      SET NEWNAME FOR DATAFILE 6 TO '?/oradata/test/indx01.dbf';
      SET NEWNAME FOR DATAFILE 7 TO '?/oradata/test/tools01.dbf';
      SET NEWNAME FOR DATAFILE 8 TO '?/oradata/test/users01.dbf';
      ALTER DATABASE RENAME FILE '/dev3/oracle/dbs/redo01.log'
          TO '?/oradata/test/redo01.log';
      ALTER DATABASE RENAME FILE '/dev3/oracle/dbs/redo02.log'
          TO '?/oradata/test/redo02.log';
    
      # Do a SET UNTIL to prevent recovery of the online logs
      SET UNTIL SCN 123456;
      # restore the database and switch the data file names
      RESTORE DATABASE;
      SWITCH DATAFILE ALL;
    
      # recover the database
      RECOVER DATABASE;
    }
    EXIT
    

12. Execute the script created in the previous step.

    For example, start RMAN to connect to the target database and run the @ command:

    % rman TARGET / NOCATALOG
    RMAN> @reco_test.rman
    

13. Open the restored database with the RESETLOGS option.

    From the RMAN prompt, open the database with the RESETLOGS option:

    ALTER DATABASE OPEN RESETLOGS;
    

    Caution:

    When you re-open your database in the next step, do not connect to the recovery catalog. Otherwise, the new database incarnation created is registered automatically in the recovery catalog, and the file names of the production database are replaced by the new file names specified in the script.

14. Optionally, delete the test database with all of its files.

    Note:

    If you used an ASM disk group, then the DROP DATABASE command is the only way to safely remove the files of the test database. If you restored to non-ASM storage then you can also use operating system commands to remove the database.

    Use the DROP DATABASE command to delete all files associated with the database automatically. The following example deletes the database files:

    STARTUP FORCE NOMOUNT PFILE='?/oradata/test/inittrgta.ora';
    DROP DATABASE;
    

    Because you did not perform the restore and recovery operation when connected to the recovery catalog, the recovery catalog contains no records for any of the restored files or the procedures performed during the test. Likewise, the control file of the trgta database is completely unaffected by the test.

In this example, the DB_UNIQUE_NAME of the primary database is MAIN and the DB_UNIQUE_NAME of the physical standby database is STANDBY. The data file sales.dbf on the primary database was lost. You want to restore this data file from the physical standby database. The service name of the physical standby database is standby_tns. The RESTORE command with the FROM SERVICE clause enables you to restore the lost data file in the primary database by using the data file in the physical standby database. The password file in the primary database and the physical standby database are the same.

Use the following steps to restore the data file sales.dbf in the primary database by using the data file in the physical standby database: