cloned db puzzle

Hi all,

After we cloned a db (8.1.7.4) to another server which has the same # CPU there is one frequently executed stored procedure that is doing 3 to 4 more times
buffer gets per execution than the original server. We checked the tables invovled
in this stored procedure, indexes, make sure they are analyzed. init parameters are the same.

I noticed on the new server the "buffer is not pinned count" is far greater than the original server. What's interesting is that on the new server, when I reviewed
statspack from the very beginning after we build it, that stored procedure is in the same ball park as the original one. As we increase load on the new server buffer gets went up, where as for the original server, the buffer gets stays the same even when load goes up.

I am looking for any suggestions. Oh, of course, the "latch free" buffer cache chain
is the top wait, which is should be the result of more buffer gets.

Thanks.

Richard

--
http://www.freelists.org/webpage/oracle-l