This document describes how to create an Oracle standby database(PCB4) from a primary database(PCB5), what should be backed up on primary database in order to enable standby database recovery, how to sync the standby database with the primary database offline archives, and how to activate standby database at last. It also mentions the daily support notes on supporting Oracle standby database.
1. Actions to be performed on primary database
Hot-backup database datafiles on the primary system.
Create standby controlfile after hot-backup done
SVRMGR > ALTER DATABASE CREATE STANDBY CONTROLFILE AS <Backup control file > ;
Ensure all online archives are archived
SVRMGR > ALTER SYSTEM ARCHIVE LOG CURRENT ;
d) Backup the current init<SID>.ora of the database.
2.Create Oracle Service for Standby database
To create a new database as standby database with same <SID> as the primary database. We could create a small one with SYSTEM tablespace initially, using Oracle Database Assistant. It helps creation of the Oracle NT services.
Use the same Oracle version on creating the database
Ensure Oracle listener could serve the new Oracle service.
On NT service, change the oracleStart<SID>?service to manually mode, to avoid it automatically start-up on the NT reboot.
Shutdown the oracleStart<SID>?service once the above steps done, and ensure the 揙racleService<SID>?service is still running on.
3.Create DB Directories
Pls create the directories for holding the database datafiles, as same as the setting as on primary database.
You should follow the same drives mapping
Pls create the directories for logs, redo logs, pfile, bdump, udump etc, similar to the primary database also.
4.Transfer files to Standby Server
Transfer all of the hot-backup datafiles to the standby system.
The file list is
backup standby controlfile
Hot-backup datafiles
All offline archives available on archived directory
backup init<SID>.ora file
The destination of the transferred files should be specified as:
Backup standby control file:
Copied to all members of the control files, the exact directory path as specified in backup
init<SID>.ora file.
Hot-backup datafiles:
Copied to original datafiles location, exactly the same directory path as on primary system.
All offline archives:
Copied to the DB archive directory, exactly the same directory path as on primary system.
Backup init<SID>.ora file:
Copied to $ORACLE_HOME/dbs or the IFILE parameter refer to location.
5.Sync Standby Database
5.1 Turn Standby DB on recovery mode
a. Ensure oracleService<SID>?service started.
b. Mount the DB
SVRMGR > CONNECT INTERNAL/ <PASSWD>
SVRMGR > STARTUP NOMOUNT PFILE=<path>/init<SID>.ora
SVRMGR > ALTER DATABASE MOUNT STANDBY DATABASE;
SVRMGR > EXIT;
We could place the above commands into a STARTUP.SQL? and it will be called
by a batch command startup_standby.bat?. See Appendix for details.
You could place a shortcut on the desktop on NT standby server, to depict the job.
5.2 Apply archived logs
a. Ensure the standby DB in recovery mode.
b. Apply the logs
SVRMGR > CONNECT INTERNAL/ <PASSWD>
SVRMGR > SET AUTORECOVERY ON;
SVRMGR > RECOVER STANDBY DATABASE;
SVRMGR > EXIT;
We could place the above commands into a RECOVER.SQL?script, and it will be called
by a batch recover.bat?. See Appendix for details.
5.3 Continuous Apply archived logs in recovery mode
Ensure a regular job to copy the primary offline archives to the standby server.
Pls see aync_log.bat?file for a reference. See Appendix for details.
You could use WINCRON to schedule a job to copy drive archives from primary site to local
archive directory with 1 hr interval.
Perform the apply archived logs in recovery mode?job once new archives arrival
Pls see sync_log.bat?file for a reference. See Appendix for details.
You could use WINCRON to schedule the job to apply archives with 1 hr interval.
Perform clean job on obsolete offline archives on standby system
Pls see archive.cmd ?file for a reference. See Appendix for details.
You could use WINCRON to schedule the job to clean obsolete archives with 2 days retention.
5.4 Activate Standby Database
To activate the standby database, copy the latest offline archive logs from primary server over to the standby system.
Apply the archives to the standby database as mentioned on Section 3.2.
Activate the standby database
SVRMGR > ALTER DATABASE ACTIVATE STANDBY DATABASE ;
Shutdown the standby DB to reset file headers and clear all buffers.
SVRMGR > SHUTDOWN NORMAL;
Open the database. You could start it by NT 揙racleStart<SID>?service and then enable the service to start automatically now.
You do not need copy any online redo logs from the primary server to the standby system once it is activated.
Take a DB hot-backup after the database opens.
5.5 Special Notes
Any physical changes that occur on the primary must be accomplished on the standby as well.
New datafiles added on primary database, file entries are added to the standby controlfile only.
If the standby database can't find the physical datafile, standby recovery will stop with an
ORA-1157 error.
Use ALTER DATABASE CREATE DATAFILE <PATH> as <PATH>; ?to recreate the physical
datafile or physically copy the missing datafiles to the standby system.
You can add or drop log files on the primary without affecting the standby DB.
If you do recreation controlfile?on the primary server, you must also recreate the controlfile for the standby database. Shut down the standby database, and use a current backup standby control file from primary server.
Jimi
GIT