Wes Harrison wrote:
> "Noons" <wizofoz2k_at_yahoo.com.au> wrote in message
> news:1125017707.850612.299530_at_g49g2000cwa.googlegroups.com...
>
>
>>Wes Harrison wrote:
>>
>>>SiSoft SANDRA Pro to do a File System Benchmark before and after the
>>>installation. Before, the drive index is about 80MB/S but this
>
> decreases by
>
>>>half after Oracle is installed. If I uninstall Oracle, the drive index
>>>returns to about 80. Reinstall Oracle and it goes back down to about
>
> 40.
>
>>Have you tried with Oracle installed but shutdown?
>>If it is started up, Oracle peeks at the redo log and control files
>>every coupla seconds. It's enough to throw disk benchmark programs
>>out the window: they assume nothing else is running.
>>
>>
>>>Does it have
>>>something to do with some of the Oracle processes interfering with
>
> SANDRA
>
>>>Pro?
>>
>>My guess: yes. Anything doing IO while SANDRA is running will
>>affect the results.
>
>
> Well I have tried it with Oracle installed but with all the services stopped
> and the performance index is about the same as though Oracle wasn't
> installed. Clearly then it has something to do with the running Oracle
> processes.
>
> What I need to know is if the Oracle processes really do cause a general
> slowdown in disk performance or whether they simply interfere with the
> SANDRA benchmark and cause it to give erroneous results.
>
> Any thoughts?
>
> Wes
The answer is far more complex than what you seem to be assuming.
I'd suggest running an AWR report along with the normal o/s tools
to determine what is actually happening.
--
Daniel A. Morgan
http://www.psoug.org
damorgan_at_x.washington.edu
(replace x with u to respond)
Received on Thu Sep 01 2005 - 11:48:41 CDT
