HansF wrote:
> On Sun, 23 Apr 2006 19:15:36 +0200, Frank van Bortel wrote:
>
>> Bob Jones wrote:
>>>>> So you have seen 2 completely identical systems, apart from the buffer
>>>>> cache, the one with 30% hits outperforms the other with 100% hits. You
>>>>> must
>>>>> have disks that are faster than memory.
>>>>>
>>>>>
>>>> You're wrong, again.
>>>> BCHR does not say a thing about performance.
>>>>
>>>>
>>> If you have seen such systems, I don't doubt you come up with this
>>> conclusion.
>> OK - let's turn this around. I give in. I have a system with a BCHR of
>> 85%. Consistently.
>>
>> Now you tell me - is the performance OK?
>
> Re-reading - in his posts Bob separates 'buffer tuning' from 'application
> tuning'.
>
> Apparently there is a need to tune the buffer cache for reasons other
> than making the application[s] work in an optimal fashion. As such, the
> BCHR might be useful.
>
> I think the bigger question to Bob is ... HOW does he use the BCHR to do
> tuning? What are the steps? ... the triggers?
>
> Most specifically, how does the use of BCHR relate to performance? Or how
> would it be used when there is a performance issue?
>
Don't agree - Bob *knows* a BCHR of 30% is bad - so he must
also know whether 85% is good or bad.
I know it means nothing, but Bob seems to think otherwise, but
cannot explain that to the 6-year old I sometimes am. Like now
--
Regards,
Frank van Bortel
Top-posting is one way to shut me up...
Received on Mon Apr 24 2006 - 12:43:44 CDT
