Chris, I suspect we've got a mix; some of the processes may be (accidentally) well served by the current settings, while others are not. Plus we've got some odd schema combinations sharing instances so I may need to move a few things around before I can test, much less make any changes. Typical stuff I imagine for systems that have been hanging around for a while. We're getting ready to go through another round of database and application upgrades and new servers are coming in, so if I can figure out what I want to try, I have a reasonable chance of getting to at least test it while we've got extra environments running. cheers ... Robyn On 7/10/07, Chris Dunscombe <chris@xxxxxxxxxxxxxxxxxxxxx> wrote:
Hi, One situation I've experienced was a smallish (< 250GB) third-party online operational database on Solaris where the OS cache acted as a cache for Full Table scans of tables around the 100-300 MB size. This worked well although it was more by accident than design. Cheers, Chris