Home » RDBMS Server » Server Administration » Use 10g to undo/check data from history (Oracle 10.2.0, Red Hat EL 5.3)
Use 10g to undo/check data from history [message #396134] Sat, 04 April 2009 02:35 Go to next message
linuxos
Messages: 31
Registered: April 2009
Member
Hi, I'm quite new in administration. When I read many documents about Flashback of 10g, I found there are many advantages of implementing it.

However, I don't know whether it will affect performance. It uses additional processes and memory to handle it.

And the DBA has to concern about the DB_FLASHBACK_RETENTION_TARGET and DB_RECOVERY_FILE_DEST_SIZE. It the transaction volume suddenly exceeds the quota, the db will stop, right? Can Oracle handle this automatically?

If the user errors do not usually happen and job rerun is possible, then can I skip this feature in the new Oracle 10g database?

Thank you!
Re: Use 10g to undo/check data from history [message #396161 is a reply to message #396134] Sat, 04 April 2009 07:35 Go to previous messageGo to next message
ebrian
Messages: 2794
Registered: April 2006
Senior Member
Flashback database is not compulsory for the database. Oracle has indicated that flashback database will produce overhead of roughly 2% (course this would depend on your database work load). DB_RECOVERY_FILE_DEST_SIZE will have to be set appropriately or your database will hang.
Re: Use 10g to undo/check data from history [message #396171 is a reply to message #396134] Sat, 04 April 2009 09:46 Go to previous messageGo to next message
linuxos
Messages: 31
Registered: April 2009
Member
Thanks!

Instead of tuning the size of flash recovery area, I think the DBA has to consider the number of redo log files and archived log files for recovery.

I've seen a case the my app suddenly perform lots of DML (insert/update/delete) in a short time, then the archive process hangs. Add enough redo logs/archived logs may help the db performace and may save DBA effort. Am I right?
Re: Use 10g to undo/check data from history [message #396174 is a reply to message #396171] Sat, 04 April 2009 10:25 Go to previous message
ebrian
Messages: 2794
Registered: April 2006
Senior Member
You should check your alert log to determine how often you are performing log switches. You may need to add additional redo logs and/or increase the size of the members you already have. Check Determine Optimal Size for Redo Logs for help with your sizing.

If you have a FRA configured and haven't set the size of the DB_RECOVERY_FILE_DEST_SIZE set appropriately and you are storing your archive logs in the FRA, then your database will hang will you reach the DB_RECOVERY_FILE_DEST_SIZE limit.
Previous Topic: Not able to get data in ascending in Oracle10g
Next Topic: User Privileges
Goto Forum:
  


Current Time: Sun Dec 04 20:38:45 CST 2016

Total time taken to generate the page: 0.07775 seconds