log switches occuring too quickly
Oracle 8 on Solaris
I have a situation where log switches are occuring too quickly (within seconds) at one site. log_checkpoint_interval is set to 0 and its not occuring in other sites with similar configuration.
Any pointers on which direction to proceed?? What should I check further?
Is your site experiencing heavy transaction which might be writing too fast into the redo log and switching.And also what is the size of your redo log files.
How big are they?
Do you have archivelogs - how big are they?
"A common mistake that people make when trying to design something completely
foolproof is to underestimate the ingenuity of complete fools" - Douglas Adams
The size of the redologs is about 10 M each (there are three) and its in Archivelog mode.
Anand, how to find whether the site is featuring 'heavy' transactions and the size of transactions? I am not getting the correct dynamic view right now.
Use the following query to know just how many transactions are taking place in your system.
select to_char(sysdate,'dd-mon-yyyy hh24:mi:ss') ttime,
where name like '%user commit%' ;
The problem seems to be solving on its own. Maybe it was a temporary phase when transaction volumes/sizes increased tremendously.
Thanks everyone for help.
Check your alert.log to see if you have any "Checkpoint not complete" messages. If not, don't worry about it. If you do have these messages, you probably need more redo for the next time you experience a heavy load.
None of those messages right now. I will keep a track, though. Thanks Jeff.
Click Here to Expand Forum to Full Width