Re: IMP/EXP accross platforms: What will be missing?

From: Rick Denoire <100.17706_at_germanynet.de>
Date: Tue, 04 Nov 2003 00:22:34 +0100
Message-ID: <3indqv4vemr6pe9ppucnn38cun2ql28f7c@4ax.com>

ydias_at_hotmail.com (dias) wrote:

>Hi,
>
>Your plan seems good, but it could take a long time to export and
>import 230 GB.
>You could try to parallelize this operation by exporting and importing
>users rather than doing just one full export and import.

Yes, good hint. On the other side, some degree of fuss may be introduced here. What about a user's grants on other users' objects? Schemas are not completely independent.

I am afraid that the limiting factor will be the network transfer rate anyway. And I am wondering if this parallel input into the DB won't physically shuffle data extents from different segments randomly, which would be bad for full table scans later (assuming the schemas being loaded in parallel reside in the same tablespace).

I can't wait to begin testing!

Bye
Rick Denoire Received on Mon Nov 03 2003 - 17:22:34 CST