PDA

View Full Version : Dbvisit and corrupted redo logs


Mike Donovan
07-02-2009, 08:43 AM
Question:
I would like know whether Dbvisit can handle corrupted redo logs?
In our environment we have a primary server, on which redo logs have accidentally been corrupted, and then have been subsequently transferred to the standby server. Dbvisit is scheduled to run every 5mins on the primary and 10 minutes on the standby. How can Dbvisit handle this situation, and what can you recommend?

Answer:
Dbvisit uses the Oracle engine to check and apply the redo logs to the standby database. Any corruption feedback will be intercepted by Dbvisit and full alerts will be sent. There are many different types of corruptions, but normally Oracle will stop and rollback to the last transaction.

Your schedule is ok, however to offer more protection (especially from human errors or software errors), you can set APPLY_DELAY_LAG_MINUTES. This creates a lag time between the archiving of a redo log at the primary database and the application of the log on the standby database. You are still 100% protected for DR, because all the logs are queued on the standby server. The logs are still transferred to the standby server without delay.

For example, if you set APPLY_DELAY_LAG_MINUTES=100, then the standby database is 100 minutes behind the primary database. So if a corruption is detected at the primary database, then it will be 100 minutes before this transaction is applied to the standby database, giving you a window to respond to this scenario.

If you need to activate the standby database for DR purposes, the lag time can be set back to 0, to ensure all transactions are applied. Dbvisit will notify you of this.