DBA Tips Archive for Oracle
No Title[an error occurred while processing this directive]
by Michael New, MichaelNew@earthlink.net, Gradation LLC
Introduced in Oracle Database 10g, Flashback Database provides users with a quick method for rewinding an entire database to a previous point in time or SCN value. This is a valuable tool used by Database Administrators to protect critical databases against user failure. Consider for example a scenario in which a user accidentally deletes several million records from a few key table on a production database. Prior to Oracle Database 10g and Flashback Database, the DBA would most likely need to perform a complex point-in-time recovery using RMAN to a point in time right before the DELETE occurred. Although a realistic solution, it has the potential for a high Mean-Time-To-Recovery (MTTR) given the database will be offline during the entire process of restoring the database's datafiles from the most recent backup followed by a point-in-time recovery of the archived redo logs that have accumulated since the last backup. If, however, the database had flashback database enabled, the DBA could avoid all of the costly and risky operations required for a point-in-time recovery operation and simply rewind the database to the same point-in-time in a fraction of the time using the FLASHBACK DATABASE command.
In addition to providing a quick and efficient method for recovering from user-induced errors, Flashback Database can also be used in combination with Oracle Data Guard, restore points, and physical standby database to temporarily open a physical standby database in read/write mode for QA testing, development fixes, or reporting purposes and then flashed back to a point in the past (the restore point) to be reverted back to a physical standby database. After the physical database is flashed back, Data Guard will automatically synchronize it with the primary which eliminates the need to formally re-create the physical standby database from a backup copy of the primary database. Throughout this entire process, the primary database can remain online without the need to log off users or shut it down.
Figure 1 provides an overview of how a physical standby database can be temporarily activated as a read/write clone of the primary database and reverted back to its role as a physical standby database using Flashback Database. This process can executed any number of times if necessary.
Figure 1: Opening Physical Standby Database in Read/Write for Testing
It is assumed that a primary and one physical standby database is already configured in order to perform the steps described in this guide. The examples used in this guide will make use of the Oracle Data Guard configuration described in the article below:
The first step is to configure the physical standby database to use flashback logging for Flashback Database operations. Although not a requirement for this exercise, it is best practice to also enable Flashback Database on the the primary.
In order to use Flashback Database, the physical database must be running in Archive Log Mode. Given the fact that this is a Data Guard configuration, both the primary and physical standby database will already be running in archive log mode and no action is required.
Flashback Database requires that the flashback logs reside in the Flash Recovery Area (FRA). Although the FRA may already be configured for your environment, the flashback logs can significantly increase FRA usage. It is not unheard of to double the size of the FRA to utilize Flashback Database.
If a FRA does not exist, it will need to be created. For example, run the following statements on the physical standby database to configure a 32GB flash recovery area in the /u03/flash_recovery_area directory with a retention time of 24 hours:
The size and location of the FRA and the Flashback Database retention policy time listed above should be modified for your environment.
Stop Redo Apply on the physical standby database and verify it is in MOUNT mode.
Turn on Flashback Logging.
Place the physical standby database back into recovery mode by starting Redo Apply.
Perform the following steps to activate the physical standby database as a production database clone which can be used for QA testing, development fixes, or reporting purposes. The clone database will be resynchronized back with the primary database and returned to its original state as a physical standby database in the next section.
Prepare the physical standby database to be activated.
On the physical standby database, stop Redo Apply and create a restore point.
When you create a guaranteed restore point, you associate an easy-to-remember name such as before_open_standby with a timestamp or SCN so that you can later flash back the database to a name instead of specifying an exact SCN or time.
Prepare the primary database to have the physical standby be diverged.
Prepare the primary database to be split from the physical standby by archiving the current log file. This step is necessary in order to write the SCN of the restore point (created in step 1) to the redo stream and archive it to the physical standby database. To accomplish this, switch logs on the primary database.
When using standby redo log files, this step is essential to ensure the database can be properly flashed back to the restore point.
On the primary database (on all instances if this is a Real Applications Cluster), defer the archival of redo data to the destination associated with the physical standby database that will be opened, and switch logs one more time to stop redo transport.
Activate the physical standby database.
On the physical standby database, perform the following steps:
Activate the physical standby database.
Skip the next statement if the physical standby has not been opened read-only since the instance was last started.
The last step is to set the protection mode to maximum performance and open the database for read/write access.
If the standby database that was opened read/write temporarily has remote archive log destinations, you should disable them. In this way, the read/write testing or reporting database will not propagate its temporary changes to other standby databases in the original Data Guard environment.
After the standby database is activated, its protection mode is downgraded to the maximum performance mode, because there is no standby database configured to protect the database against data loss while it is temporarily activated as a production database. Note that this protection mode setting does not affect the protection mode of the original primary database, it affects only the activated standby database.
When the activated standby database is converted back to a physical standby database, its protection mode is automatically changed to match that of the original primary database.
Once the physical standby database has been activated, it is a full-blown clone of the primary database. You may perform QA testing, run reports, fix bugs, test new code, or create objects on the new activated database for as long as needed, independent of the primary database.
While the database is activated, it is not receiving redo data from the primary database and cannot provide disaster protection. It is recommended that there be at least two physical standby databases participating in the configuration so that the primary database remains protected against data loss.
Also, any results stored in the activated database will be lost when you later flash back the database to before the activation time. Results that should be saved must be copied or exported out of the activated database before flashing it back.
After testing is completed, you need to resynchronize the activated database with the primary database. Issue the following statements on the activated database to quickly flash it back to the guaranteed restore point created earlier (before_open_standby) and resynchronize it with the primary database.
Revert Physical Standby Database using Flashback Database.
Catch up the standby database to the primary database.
The method you use in this step will depend on how far the activated standby database lags behind the primary database in its application of redo data.
Let archive gap resolution fetch all missing archived redo log files and allow Redo Apply to apply the gap.
If the activated database has not fallen too far behind the original primary database, issue the following statement on the standby database to resynchronize it with the primary database and restart Redo Apply. For example:
Then, go to step 3.
Create an incremental backup on the primary and apply it to the standby.
If the activated database has fallen too far behind the original primary database (for example, if there are not sufficient log files available), you can take an incremental backup from the primary database and apply it to the standby database.
The following guide provides the steps necessary on use an RMAN incremental backup to roll forward and resynchronize a physical standby database that lags far behind the primary database:
After you apply an incremental backup to the standby database, you typically need to apply more redo to the standby database to activate the physical standby database again for read/write testing or reporting purposes. More specifically, you might need to apply the redo generated by the primary database while the incremental backup was taken. Otherwise, issuing an ALTER DATABASE ACTIVATE STANDBY DATABASE will return an error.
Re-enable archiving to the physical standby database destination.
On the primary database, issue the following statement to re-enable archiving to the physical standby database:
Jeffrey Hunter is an Oracle Certified Professional, Java Development Certified Professional, Author, and an Oracle ACE. Jeff currently works as a Senior Database Administrator for The DBA Zone, Inc. located in Pittsburgh, Pennsylvania. His work includes advanced performance tuning, Java and PL/SQL programming, developing high availability solutions, capacity planning, database security, and physical / logical database design in a UNIX / Linux server environment. Jeff's other interests include mathematical encryption theory, tutoring advanced mathematics, programming language processors (compilers and interpreters) in Java and C, LDAP, writing web-based database administration tools, and of course Linux. He has been a Sr. Database Administrator and Software Engineer for over 20 years and maintains his own website site at: http://www.iDevelopment.info. Jeff graduated from Stanislaus State University in Turlock, California, with a Bachelor's degree in Computer Science and Mathematics.
Copyright (c) 1998-2018 Jeffrey M. Hunter. All rights reserved.
All articles, scripts and material located at the Internet address of http://www.idevelopment.info is the copyright of Jeffrey M. Hunter and is protected under copyright laws of the United States. This document may not be hosted on any other site without my express, prior, written permission. Application to host any of the material elsewhere can be made by contacting me at firstname.lastname@example.org.
I have made every effort and taken great care in making sure that the material included on my web site is technically accurate, but I disclaim any and all responsibility for any loss, damage or destruction of data or any other property which may arise from relying on it. I will in no case be liable for any monetary damages arising from such loss, damage or destruction.
Last modified on
Thursday, 17-Apr-2014 02:54:07 EDT
Page Count: 42399