Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Snapshot Too Old Reported - But We Doubt It

Re: Snapshot Too Old Reported - But We Doubt It

From: Alan <alan_at_erols.com>
Date: Wed, 4 Aug 2004 10:13:41 -0400
Message-ID: <2nc98gFvi12sU1@uni-berlin.de>

"Joel Garry" <joel-garry_at_home.com> wrote in message news:91884734.0408031413.6b87428a_at_posting.google.com...
> "Alan" <alan_at_erols.com> wrote in message
news:<2n7ak8Fr1bvqU1_at_uni-berlin.de>...
> >
> > We searched Metalink, but nothing we could find addressed the problem.
> >
> > Any ideas what is wrong and how to solve it? TIA.
>
> Well, Howard's explanation is better than the ones on metalink, but
> there certainly are enough "what to do about 1555" docs there. Here's
> one that refers to the old ones and explains an undo iteration of it:
>

http://metalink.oracle.com/metalink/plsql/ml2_documents.showFrameDocument?p_database_id=NOT&p_id=269814.1
>
> The only biggie I haven't seen mentioned in this thread is "are you
> using OPTIMAL?" Don't use OPTIMAL.
>

Thanks for the reply. I wound up taking another approach to the problem, as even with using a large RB segment, we still got the 1555. Reading in Tom Kyte's book, it seems this can happen even if your RBS is big enough, which it was, and it did.

Anyway, I wound up doing a CTAS on just the subset of data that I needed from the larger table, and then using the new table to update the values in the table that needed updating. Much faster, and no 1555. I had violated one of my own rules, which is: Don't be afraid to break up a large transaction into smaller ones. Received on Wed Aug 04 2004 - 09:13:41 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US