Re: Imported DB runs SLOWER on faster machine

Sorry Rick

Norman and I may be side tracking you. The default behaviour of imp is to estimate statistics on import, which may or may not be good enough for you, it sounds like it isn't. I was replying to Vladimir who had said you don't get stats on import, and illustrating it by the error I get on a particular import job of mine. Norman was pointing out that my problem could be avoided by changing the statistics parameter used by imp.

As for your issue, assuming all non-os parameters are the same I'd start with analysing the db on the NT box in the same way as the box from whence it came. You should get similar performance after this step.

The other thing I can't find any of us asking you to do, and you should do it after collecting stats if you still have the problem, is to collect an explain plan for the long running query on both boxes. In an ideal world the execution plans would be the same.

Finally you also asked about datafiles. These have a limit to the number of blocks that they can contain. as you have > 20gb datafiles the likely limit for you is 32gb (8k block size * 2^32 blocks)l if your block size is 16k then the max size will be correspondingly larger.

--
Niall Litchfield
Oracle DBA
Audit Commission UK
*****************************************
Please include version and platform
and SQL where applicable
It makes life easier and increases the
likelihood of a good answer
******************************************
"Rick Denoire" <100.17706_at_germanynet.de> wrote in message
news:vrgrju8rsrh6ob2qrrfir1hrn61mhmrp2l_at_4ax.com...


> Norman Dunbar <Norman.Dunbar_at_lfs.co.uk> wrote:

>

>

> >

> >have you rtried setting STATISTICS=NONE on the export ?

> >

>

> No, I haven't. As a matter of fact, I did not know about this option.

>

> I will go after that. Still learning.

>

> Rick