Re: AWR interpretation

From: Andy Klock <>
Date: Wed, 30 Oct 2013 16:10:12 -0400
Message-ID: <>

On Wed, Oct 30, 2013 at 3:52 PM, Storey, Robert (DCSO) <> wrote:
> I have to check but I'm pretty sure one table has 3400 records that have lob (pictures) in them.
> From: Andrew Kerber []
> Sent: Wednesday, October 30, 2013 2:51 PM
> To: David Barbour
> Cc: Storey, Robert (DCSO); Oracle L
> Subject: Re: AWR interpretation
> One other thing, if you have lob's in your data, you can pretty much kiss your performance goodby. I am sure someone here can identify why, I never really tried to track it down, but basically lob's take about 10 times longer to deal with than any other item.

Tables that have LOBs can't do array inserts, so it's a one row at a time sort of thing. Plus, the rows tend to be bigger and touch lots of blocks.

You mentioned parallizing the import, this can possibly be done if the 46 million row table is partitioned. Otherwise, the export for that table can be split up with QUERY (exp help=y, I have no idea what the OEM database control export interface looks like, so I can't offer any guidance there. But I would first test to see if the reduction of COMMITs is all that is needed.

Received on Wed Oct 30 2013 - 21:10:12 CET

Original text of this message