While reading a 10.2.0.3 statspack report I was wondering why the
authors chose to compute the "Library Hit %" from the PINHITS rather
than from the GETHITS of v$librarycache.
In the report I've been reading, the lack of bind variables can be
guessed from the 100% Miss for the "Get Requests" of the SQL AREA
namespace in the section "Library Cache Activity". Reading the
"Library Hit %" at the beginning of the report, it is less obvious:
90%.
Of course ratios are never relevant without further investigation. But
could anyone tell me why PINHITS was apparently deemed more important
than the GETHITS by statspack authors when computing the hit ratio?
Received on Fri Sep 14 2007 - 10:43:33 CDT
