Today I am going to cover Dataguard which is used for business continuity for Oracle Applications 11i including fusion middleware (Using AS Guard). There are two types of standby databases Physical Standby Database & Logical Standby database. Since oracle applications currently support only physical standby database , I am going to cover only physical standby databases.
Overview of Dataguard & Dataguard Broker
Dataguard consist of one primary database & one or more standby database where primary & standby databases are connected by sqlnet (How to configure this check in coming posts). These primary & standby database can be monitored by command line (sql) as well from GUI (Graphical User Interface) Tool called as Dataguard broker.
Important Services in Dataguard configuration
Two very important services in dataguard configuration are Redo Transport Service & Log Apply Services. In order to keep standby database in synch with primary database ; redo transport service transfer(Pull/Push) redo logs from primary to standby database & Log apply service applies that redo logs to standby database.
Switchover & Failover
These are basically two kind of roles in dataguard Failover & Switchover; In Switchover Primary database becomes standby database & one of (If there are more than one standby) standby database becomes primary database for ex. during server maintenance. Failover is used in scenarion in which your primary database is not available and you want to convert role of one of standby database to primary database for ex. during primary database failure.
Data Protection Mode
Standby database in dataguard configuration can be configured in any of three protection mode which can be changed later as well depending on your requirement . These protection modes are Maximum Protection Mode , Maximum Availability Mode & Maximum Performance Mode.
Maximum Protection Mode
In maximum protection mode as name suggest no data loss will occur if primary database fails. This is achieved by oracle by ensuring that redo data must be written to both local online redo log location & standby redo log location before transaction commits. This standby redo log will ensure full data protection. For full protection mode atleast one standby database should have standby redo log and LGWR, SYNC, and AFFIRM attribute should be used with LOG_ARCHIVE_DEST_n parameter. (I am going to discuss about these parameters in next post)
Here your primary database will shutdown if it is unable to write redo records to atleast one standby location.
Maximum Availability Mode
This is similar to Maximum protection mode except that it will not shutdown primary database if primary database is not able to write standby redo logs to atleast one standby location but in that case Maximum Availability will act in Maximum Performance Mode until fault is corrected and all gaps in redo logs are resolved. Once all gaps are resolved primary database will automatically goes into Maximum Availability Mode.
Maximum Performance Mode
This is default mode in which redo stream is written asynchronously with respect to the commitment of the transactions that create the redo data. You can use LGWR and ASYNC attribute or ARCH attribute in LOG_ARCHIVE_DEST_n
You can check your database mode as
SQL> select PROTECTION_LEVEL from v$database;
How to change protection modes..
How many standby database Instance you can configure ...
explanation on Log apply, log transport service , FAL Server/Client ....
Step by step configuration of standby database using dataguard ..