----- Original Message -----
Sent: Tuesday, September 10, 2002 4:23
AM
Subject: Nasty bug in 9.2.0.1 (was:
Oracle and Cache Hit Ratio based train
FWIW
There is a document (id 283852) on Metalink updated on
03-SEP-2002 created by "CENTER OF EXPERTISE" named
THE COE PERFORMANCE METHOD
A
PERFORMANCE METHODOLOGY FOR ENTERPRISE-WIDE INFORMATION SYSTEMS
which shows that Oracle Corporation turned from "ratio" to
"wait interface" approach.
The bad news is that in version 9.2.0.1 view V$SESSION_EVENT
broken (bug 2429929), also described in Notes 208066.1. Bug fixed in
10i.
This bug affects OEM, BSTAT/ESTAT, STATSPACK, etc.
Did anybody adjust scripts which join V$SESSION with
V$SESSION_EVENT?
Should we use something like
s.sid=e.sid+1 or s.sid-1=e.sid?
Alex.
-----Original Message-----
From:
Johnson, Michael [mailto:[EMAIL PROTECTED]]
Looks like Cache Hit Ratios may finally go away !!!
