And I think it's important to realise that ratios are useless as a
starting point in the tuning process on any system, not only Oracle.
Most OS'es and databases use not instrumented correctly to deal with
response time measurements (makes you wonder: If response TIME is what
matters, how can
Yep.
Simple example: Even though it seems to be sometimes a
little on the 'random' side, the ELAPSED_TIME column
on V$SQL in v9 is an absolute god send...
Cheers
Connor
--- Mogens_Nørgaard [EMAIL PROTECTED] wrote: And I
think it's important to realise that ratios
are useless as a
starting
I think Dave Ensor, at the recent UKOUG conference, called it his
portable tuning kit:
select elapsed_time, cpu_time
from v$sql
order by elapsed_time;
Mogens
Connor McDonald wrote:
Yep.
Simple example: Even though it seems to be sometimes a
little on the 'random' side, the ELAPSED_TIME
Hi Yong
I agree that It's not the ratio that needs condemning, it's
the advice about... What I disagree is the wrong educational
tool people on public forums have recently used again and
again to show the inadequacy of the BCHR tuning method.
The thing is most people who have used this
#1. what is the difference in the 4 BCHR's in 9i?
#2. How do you determine whether your buffer cache is sized properly?
- Original Message -
To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]
Sent: Saturday, December 27, 2003 9:34 AM
Hi Yong
I agree that It's not the ratio
#1. what is the difference in the 4 BCHR's in 9i?
What do you mean by that? Are you refferring to different buffer pool stats
here?
#2. How do you determine whether your buffer cache is sized properly?
Your application meets it's defined operational response time and throughput
constraints.
#2. How do you determine whether your buffer cache is sized
properly?
Your application meets it's defined operational response time
and throughput constraints.
Actually that is an interesting answer, since it doesn't answer the question
asked 'How do I know if my buffer cache is sized
Performance 'problems' are dependent on what the marketing department gets
in the SLA.
So if your marketing guys negotiate very strict response time requirements
and you dont meet them, then you have a performance problem. How do I know
if my buffer cache is having any effect on that?
-
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Ryan
Sent: 27 December 2003 21:59
To: Multiple recipients of list ORACLE-L
Subject: Re: Correct way to accuse BCHR tuning method (Was: Hit ratio)
Performance 'problems' are dependent on what
#2. How do you determine whether your buffer cache is sized
properly?
Your application meets it's defined operational response time
and throughput constraints.
Actually that is an interesting answer, since it doesn't answer the
question
asked 'How do I know if my buffer cache is
Yong,
The point I want to make is that the buffer cache hit ratio actually
*is* useful, in the following way. If its value is greater than about
99%, then I can practically guarantee that there is some high-LIO SQL
inside the application whose repair will greatly improve system
performance.
BCHR tuning is useless as a starting point in the tuning process.
Anjo.
-Original Message-
Yong Huang
Sent: Wednesday, December 24, 2003 6:09 PM
To: Multiple recipients of list ORACLE-L
[This message is not technical, but educational. Readers interested in
technical info only may want
12 matches
Mail list logo