Re: Cache Hit Ratio from system views

On Aug 16, 8:28 pm, "Bob Jones" <em..._at_me.not> wrote:
> >> Why is BHCR meaningless? The answer should be short and simple. I want
> >> to hear your opinion.
>
> > One can not prove a negative.
> > Where is your proof BCHR is a reliable indicator of GOOD performance?
>
> BCHR alone does not tell you about overall performance. It simply tell you
> the disk I/O percentage. It is an indicator, a very meaningful one.

Others have shown why it is not a meaningful indicator. I would just like to add the observation for those who might not already know, this is one of the prime examples of how oversimplifying a metric can leach into the education of Oracle, becoming enshrined in performance tuning scripts and propagated into myth. It is worse than meaningless, it is actually misleading when people think it is showing disk I/O percentage. It may have some corrrelation to such a percentage on a very simple system with a very simple test (and can just as simply be bogosified), but most production systems actually have multiple users sharing and updating data, with hardware that may not be _telling the truth to Oracle_ as to whether it is actually doing I/O. Throw an SSD in there and it is all lies.

I don't have any argument with the idea that response times meeting business requirements is a, or even the most reasonable metric. But I feel compelled to point out how common it is to not define those requirements properly, so making an assumption that there are already properly defined response time requirements is jumping ahead too far in a performance methodology. And yet, by the time there is a perceived problem, it is already too late.

jg

--
@home.com is bogus.
Isn't it ironic...they didn't even deal with the subrime market...
http://www.firstmagnus.com//