为什么80%的码农都做不了架构师？>>> 
Recovery manager is a platform independent utility for coordinating your backup and restoration procedures across multiple servers. In my opinion it's value is limited if you only have one or two instances, but it comes into it's own where large numbers of instances on multiple platforms are used. The reporting features alone mean that you should never find yourself in a position where your data is in danger due to failed backups.
The functionality of RMAN is too diverse to be covered in this article so I shall focus on the basic backup and recovery functionality.
Create Recovery Catalog
Register Database
Full Backup
Restore & Recover The Whole Database
Restore & Recover A Subset Of The Database
Incomplete Recovery
Disaster Recovery
Lists And Reports
Create Recovery Catalog
First create a user to hold the recovery catalog:
CONNECT sys/password@w2k1 AS SYSDBA
-- Create tablepsace to hold repository
CREATE TABLESPACE "RMAN"
DATAFILE 'C:\ORACLE\ORADATA\W2K1\RMAN01.DBF' SIZE 6208K REUSE
AUTOEXTEND ON NEXT 64K MAXSIZE 32767M
EXTENT MANAGEMENT LOCAL
SEGMENT SPACE MANAGEMENT AUTO;
-- Create rman schema owner
CREATE USER rman IDENTIFIED BY rman
TEMPORARY TABLESPACE temp
DEFAULT TABLESPACE rman
QUOTA UNLIMITED ON rman;
GRANT connect, resource, recovery_catalog_owner TO rman;
Then create the recovery catalog:
C:>rman catalog=rman/rman@w2k1
Recovery Manager: Release 9.2.0.1.0 - Production
Copyright (c) 1995, 2002, Oracle Corporation. All rights reserved.
connected to recovery catalog database
recovery catalog is not installed
RMAN> create catalog tablespace "RMAN";
recovery catalog created
RMAN> exit
Recovery Manager complete.
C:>
Register Database
Each database to be backed up by RMAN must be registered:
C:>rman catalog=rman/rman@w2k1 target=sys/password@w2k2
Recovery Manager: Release 9.2.0.1.0 - Production
Copyright (c) 1995, 2002, Oracle Corporation. All rights reserved.
connected to target database: W2K2 (DBID=1371963417)
connected to recovery catalog database
RMAN> register database;
database registered in recovery catalog
starting full resync of recovery catalog
full resync complete
RMAN>
Existing user-created backups can be added to the catalog using:
RMAN> catalog datafilecopy 'C:\Oracle\Oradata\TSH1.dbf';
RMAN> catalog archivelog 'log1', 'log2', 'log3', ... 'logN';
Full Backup
First we configure several persistant parameters for this instance:
RMAN> configure retention policy to recovery window of 7 days;
RMAN> configure default device type to disk;
RMAN> configure controlfile autobackup on;
RMAN> configure channel device type disk format 'C:\Oracle\Admin\W2K2\Backup%d_DB_%u_%s_%p';
Next we perform a complete database backup using a single command:
RMAN> run {
2> backup database plus archivelog;
3> delete noprompt obsolete;
4> }
The recovery catalog should be resyncronized on a regular basis so that changes to the database structure and presence of new archive logs is recorded. Some commands perform partial and full resyncs implicitly, but if you are in doubt you can perform a full resync using the follwoing command:
RMAN> resync catalog;
Restore & Recover The Whole Database
If the controlfiles and online redo logs are still present a whole database recovery can be achieved by running the following script:
run {
shutdown immediate; # use abort if this fails
startup mount;
restore database;
recover database;
alter database open;
}
This will result in all datafiles being restored then recovered. RMAN will apply archive logs as necessary until the recovery is complete. At that point the database is opened. If the tempfiles are still present you can issue a command like like the following for each of them:
sql "ALTER TABLESPACE temp ADD
TEMPFILE ''C:\Oracle\oradata\W2K2\temp01.dbf''
REUSE";
If the tempfiles are missing they must be recreated as follows:
sql "ALTER TABLESPACE temp ADD
TEMPFILE ''C:\Oracle\oradata\W2K2\temp01.dbf''
SIZE 100M
AUTOEXTEND ON NEXT 64K";
Restore & Recover A Subset Of The Database
A subset of the database can be restored in a similar fashion:
run {
sql 'ALTER TABLESPACE users OFFLINE IMMEDIATE';
restore tablespace users;
recover tablespace users;
sql 'ALTER TABLESPACE users ONLINE';
}
Incomplete Recovery
As you would expect, RMAN allows incomplete recovery to a specified time, SCN or sequence number:
run {
shutdown immediate;
startup mount;
set until time 'Nov 15 2000 09:00:00';
# set until scn 1000; # alternatively, you can specify SCN
# set until sequence 9923; # alternatively, you can specify log sequence number
restore database;
recover database;
alter database open resetlogs;
}
The incomplete recovery requires the database to be opened using the RESETLOGS option.
Disaster Recovery
In a disaster situation where all files are lost you can only recover to the last SCN in the archived redo logs. Beyond this point the recovery would have to make reference to the online redo logs which are not present. Disaster recovery is therefore a type of incomplete recovery. To perform disaster recovery connect to RMAN:
C:>rman catalog=rman/rman@w2k1 target=sys/password@w2k2
Once in RMAN do the following:
startup nomount;
restore controlfile;
alter database mount;
From SQL*Plus as SYS get the last archived SCN using:
SQL> SELECT archivelog_change#-1 FROM v$database;
ARCHIVELOG_CHANGE#-1
--------------------
1048438
1 row selected.
SQL>
Back in RMAN do the following:
run {
set until scn 1048438;
restore database;
recover database;
alter database open resetlogs;
}
If the "until scn" were not set the following type of error would be produced once a redo log was referenced:
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of recover command at 03/18/2003 09:33:19
RMAN-06045: media recovery requesting unknown log: thread 1 scn 1048439
With the database open all missing tempfiles must be replaced:
sql "ALTER TABLESPACE temp ADD
TEMPFILE ''C:\Oracle\oradata\W2K2\temp01.dbf''
SIZE 100M
AUTOEXTEND ON NEXT 64K";
Once the database is fully recovered a new backup should be perfomed.
The recovered database will be registered in the catalog as a new incarnation. The current incarnation can be listed and altered using the following commands:
list incarnation;
reset database to incarnation x;
Lists And Reports
RMAN has extensive listing and reporting functionality allowing you to monitor you backups and maintain the recovery catalog. Here are a few useful commands:
# Show all backup details
list backup;
# Show items that beed 7 days worth of
# archivelogs to recover completely
report need backup days = 7 database;
# Show/Delete items not needed for recovery
report obsolete;
delete obsolete;
# Show/Delete items not needed for point-in-time
# recovery within the last week
report obsolete recovery window of 7 days;
delete obsolete recovery window of 7 days;
# Show/Delete items with more than 2 newer copies available
report obsolete redundancy = 2 device type disk;
delete obsolete redundancy = 2 device type disk;
# Show datafiles that connot currently be recovered
report unrecoverable database;
report unrecoverable tablespace 'USERS';
It's worth spending some time looking at all the reporting capabilities whilst deciding whether you should switch from shell scripting to RMAN. It might just influence your decision.