The default algorithm of ASSM is that memory is stolen from the buffer cache if the shared pool is stressed.
However this default behaviour can be changed by setting _memory_broker_shrink_heaps=30
The default value for the above parameter is 0
If you want to continue using this feature you might want to set a minimum value for your buffer cache so that memory does not shrink to below this value.
Richard Saints wrote:
Hello to you all,
I've a mobile application that uses Oracle Lite 10g ( 10.2.0.2) as mobile server and Oracle Database Server 10g release 1 (10.1.0.2) as the repository database. Both of these components are installed on servers with Microsoft Windows 2000 Advanced Server.
My question is related with the repository database:
We use Automatic Shared Memory Management and have a SGA target = to the SGA Max Size = 1336Mb. ASMM always gave a bigger slice of memory to the Buffer Cache (about 70%), but in the last 2 weeks I noticed that the value of the Shared Pool started to increase and now it's bigger than Buffer Cache (a slice of 49,7%).
Here's my current situation:
Shared Pool 664 Mb
Buffer Cache 648 Mb
Large Pool 8 Mb
Java Pool 8 Mb
Other 8 Mb
I have some knowledge of Oracle Database Server Architecture, but I'm being able to interpret this behavior or why this is happening. Can anyone give me some help or some ideas ?
It's also important to say that we have been experiencing some performance problems in the last two weeks. We are also having a bigger volume of data.
Thanks for all your attention.
Senior Oracle Consultant
M: +61 402 792 405
Luggage? GPS? Comic books?
Check out fitting gifts for grads at Yahoo! Search.