I have my redo logs set to 2048MB and I am still switching between 30 to 45 minutes. But, my load will double within a year, so my switching times should shrink.
Printable View
I have my redo logs set to 2048MB and I am still switching between 30 to 45 minutes. But, my load will double within a year, so my switching times should shrink.
That's some mega redo logs.
Point of interest:
If you expect your load to double, and currently you switch every 35 - 45 minutes. Is this a case of over proactive tuning... I mean, I is up to you of course, how often you want to switch (ahem, recover)... or would it be better to decrease the rollback size to optimal swtich to 15 minutes as per Oracle recommendation.
I'm not having a go at you in this regard, I have 250 MB redo logs that switch every 2 hours, but I also expect the database to grow dramatically.
It's just a thought, it is a proactive process or incorrect administration... hmmm.
Cheers,
I am in a different boat than your average DBA. I work on a development team and also support operations. Our database needs to be fully-automated, as possible, almost like its an embedded database. So, I have to trade-off certain things to ensure operations and overly-large redo logs is one of them.
Thank you for asking.