Re: CURSOR_SHARING survey
Date: Wed, 23 Jan 2013 09:16:16 +0100
Message-ID: <CAJu8R6h4oztBsQiNetxGYhO3DO8hYwOf0QnS563KcrDXQMzUaw_at_mail.gmail.com>
Dears,
I have been confronted to a situation where setting the cursor_sharing parameter to FORCE worked as a temporary work around. A summary of this situation can be found here:
http://hourim.wordpress.com/2011/06/16/bind-variable-shared-pool-and-cursor-sharing-parameter/
Best Regards
Mohamed Houri
2013/1/22 Stojan Veselinovski <stojan.veselinovski_at_gmail.com>
> Hi Chris,
>
> As per recommendations(?) from EMC we have it set in 1 or our
> databases. A DBA out in the the business set it on the fly in a
> another prod environment believing he had a parsing problem (different
> application) with pretty dire consequences.
>
> These are the only 2 databases out of 50+ that have had it set to
> FORCE at any one point in time.
>
> I'm just starting on a performance tuning project which is looking at
> the custom interface developed for EMC Documentum and the various
> issues they are having around it. The product uses its own version of
> SQL called DQL - hence the cursor_sharing=force. Going to be a
> challenge to get this running well. Also, from what I can see early on
> most fields are made not null and a whitespace is added when they
> should be null : )
>
> To be fair to the product, at this stage I'm not sure if it's an issue
> with Custom developed code or the base product and I'm not even sure
> if the cursor_sharing=force is specified in their documentation or was
> a recommendation by the people that set it up.
>
> Stojan.
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>
-- Bien Respectueusement Mohamed Houri -- http://www.freelists.org/webpage/oracle-lReceived on Wed Jan 23 2013 - 09:16:16 CET