This topic describes the procedures carried out from the tlmenu Configuration and Setup menu to configure Rapid Recovery on a Reality database.
To use Rapid Recovery on a particular database it must first be configured. This is carried out from the tlmenu Configuration and Setup menu, displayed by selecting option 2 on the main administration menu. Refer to tlmenu for an overview.
At the Configuration and Setup menu, select option 1 to initiate the procedure to configure the database for Rapid Recovery.
Transaction Logging Menu System Wed Sep 7 14:14:33 2005 Database name : dbase2 Host name : crime State : Transaction Handling/Logging enabled Configuration and Setup ======================= 1. Define/Redefine the Database Configuration 2. Start Transaction Handling/Logging 3. Stop Transaction Handling/Logging 4. Configure the Transaction Logging Status Monitor 5. Configure backup method Enter option (1-5) :
Transaction Logging Configuration and Setup Menu
y
at the confirmation prompt.You are asked to select the resilience option that you want to configure, from the menu:
Configure database for :
1. Transaction Handling
2. Stand-Alone Transaction Logging
3. Shadow
4. FailSafe/Heartbeat (Heartbeat on UNIX Only)
5 Stand-Alone Rapid Recovery
Enter option (1-5) :
Either:
Select option 2 to configure both Rapid Recovery and Transaction Logging. When prompted:
Do you wish to set up database for Rapid Recovery File System? (y/n/q) ?
Enter y and follow the procedure for
Configuring a Database for
Transaction Logging.
Or:
Note: Transaction Handling must be running in order for Rapid Recovery to be possible. If logging is stopped on the database – either by executing TL-STOP at TCL or via option 3 on the tlmenu Configuration and Setup Menu – the Rapid Recovery feature is inhibited. It will not be possible to recover the database following a system failure until Transaction Handling is restarted.
When a database is configured for Rapid Recovery with Transaction Logging, individual files can have three levels of resilience:
Logged files (marked DL in their file definition item) have all updates logged in a clean log for auditing and disaster recovery purposes. These files are fully recoverable after a system crash. This mode should be used for main data files.
Note: Recoverable files (marked DR in their file definition item) have recent updates logged on the raw log only. These files are fully recoverable after a system crash. This mode can be used for files that are derived from main data files, for example, application level indexes. When you create an index with the CREATE-INDEX command, it is automatically set to be Recoverable.
Recoverable files are not transferred to a Failsafe secondary and hence should not be used in a Failsafe pair.
Scratch files (marked D in their file definition item) have only recent structural changes recorded in the raw log. After recovery, these files will be structurally intact but may not contain valid data. This mode can be used for scratch files, which are typically cleared when a user logs on.
When a database is configured for Rapid Recovery without Transaction Logging, individual files have two levels of resilience:
Recoverable files (marked DL or DR in their file definition item) have recent updates logged on the raw log only. These files are fully recoverable after a system crash.
Scratch files (marked D in their file definition item) only have recent structural changes recorded in the raw log. After recovery, these files will be structurally intact but may not contain valid data. This mode can be used for scratch files, which are typically cleared when a user logs on.
To set the logging status of files, use the TL-SET-LOG-STATUS command.