On Sat, 31 May 2003 23:22:03 +0000, Mladen Gogala wrote:
> On Sat, 31 May 2003 22:35:13 +0000, Holger Marzen wrote:
>
>> We run a customer's application on Oracle 8.1.7. From time to time
>> it's not possible to log in anymore. The shared seems to be
>> fragmented. A restart of the instance helps, but then the
>> application has to be restarted as well.
>>
>> Is it possible to de-fragment the shared pool or at least clear it
>> while the instance is up and the application stays connected?
>
> Have you tried "alter system flush shared_pool;"? It's an emergency
> solution only, but can make your instance survive until the next
> maintenance.
Did you check on metalink for bugs related to Shared Pool ? You might
want to consider recommending the customer to patch to 8174 . On the
other hand , check x$ksmsp every a few minutes to see if free space is
decreasing and also make sure if you are ending up with many smaller
chunks of free space . All is worth if you see any 4031s in the alert
log .
Received on Sun Jun 01 2003 - 01:58:10 CDT