Re: ORa error 03113

On Sat, 9 Mar 2002 15:01:09 +0100, "Kasper van Beem" <kbeem_at_xs4all.nl> wrote:

>Hello Sybrand,
>
>Thx, for the reply, I see I must get some things straightend out.Here are
>some more details.
>
>you're right about the hardware, the test server has only 1GB of memory
>while the production server has 1,5GB of memory.
>The test server has only one NIC while the production has two. It isn't a
>problem for the listener, I think, cause that one binds to localhost, and
>the handlers work alright. Further more the second NIC isn't used by the
>clients to connect to... Or do I make a erronous thought?
>Does it help to reduce settings like shared pool size to get rid of the
>problem or will I only make it worse if I do that?
>The real question is: Is it worth it to make the effort? Or must I "just"
>add more memonry and a NIC to resolv the problem?
>
>Monday, I will go back to work, and try to get the extra hardware.
>
>Grtz,
>
>Kasper
>
>

I don't think the NIC has anything to do with it. ORA-3113 only arises after communication has been established, and the server process suddenly died.
Reducing the shared pool will decrease the size of the shared memory segment Oracle needs to create when starting. If the problem is related to the size of the SGA there will be error messages in the alert_${ORACLE_SID}.log located in the directory pointed to be the background_dump_dest parameter in $ORACLE_HOME/dbs/init${ORACLE_SID}.ora. The default location on Unixes for the alert is IIRC $ORACLE_HOME/rdbms/log. Reducing the shared_pool can resolve the problem, it will also impair performance, assuming it hasn't be thoughtlessly set sky high without reason.

I would go for the memory, generally speaking it is a good idea a fallback system has the same specs as the production system. The NIC is as far as I can see from here unimportant.

Regards

Sybrand Bakker, Senior Oracle DBA

To reply remove -verwijderdit from my e-mail address Received on Sat Mar 09 2002 - 09:45:21 CST