Re: Snapshot too old but UNDO_RETENTION very high
Date: Fri, 6 Feb 2004 21:53:44 -0800
Message-ID: <m4mdnZeJQdNp5rndRVn-tw_at_comcast.com>
[Quoted] Hello Sybrand,
I believe we disagreed on very different issues:
In case you are discussing (lengthy selects for example) , high transaction activity from other sessions would definitely cause ora-1555 at some point
In a case, that user describes, looks like lengthy update is failing. My recommendation is to break one big update into small batched updates using PL/SQL (or any other language - Perl, fro example). This will greatly reduce probability of 1555 on an update itself.
Overall, let user decide which advice helps to resolve the problem and not to flame each other.
Regards,
Support
DBA Infopower
The advice provided by DBA Infopower are provided "as is" with no warranty. DBA Infopower expressly disclaims any warranty, regarding the advice including any implied warranty of merchant fitness for a particular purpose aviating course of dealing and/or performance. DBA Infopower does not warrant that the advices provided by DBA Infopower will be free from bias, detests, errors, eavesdropping or listening. DBA Infopower shall not be responsible for the quality of information or the authentication of the services or details given by experts on the advice. By using this advice the user of the advice accepts the terms and conditions of this statement.
"Sybrand Bakker" <gooiditweg_at_nospam.demon.nl> wrote in message
news:lur720dm952kvtnohjn6botk3i6jd1ni5p_at_4ax.com...
> On Fri, 6 Feb 2004 11:03:29 -0800, "DBA Infopower Support"
> <support_at_dbainfopower.com> wrote:
>
> > Frequent commits would greatly reduce probability of ORA-1555,
since
> >consistent data required by transaction would have much less chance to be
> >lost in rollback segments.
>
> You are just plain and utterly wrong, andapparently you don't know at
> all how the rollback mechanism works.
> Please read up on the causes of ora-1555 and apologize in this forum
> for your misguided advice.
> As far as your assesment of Metalink: I don't think you know what you
> are talking about, you are just trying to put people sand in their
> eyes. The causes for ora-1555 are well known, and they are published,
> apart from that, Metalink is currently staffed by trained monkeys
> providing similar answers like you do.
>
>
> --
> Sybrand Bakker, Senior Oracle DBA
Received on Sat Feb 07 2004 - 06:53:44 CET