All,
I have a vendor application where I CANNOT CHANGE THE CODE!!! They
recently made a change in the app (critical functionality) and have a
script that is running over 900-1500 times per second during peak
periods. Until the vendor produces a fix for this, I have to live with
it. This is a large hospital application running three hospitals, so we
are living with it for now.

Currently, the objects that this select statement accesses are
relatively small. I have a 36G SGA with 5G allocated to the keep pool.
I am thinking of altering these objects from the default cache to the
keep pool, and wanted to get some opinions on this strategy.

My hope here is to isolate this poor behavior from the default pool.
Any thoughts/opinions would be great.

Nick

Oracle LazyDBA home page