Re: Same Query, same database, same application, new plan from one environment but not from another
Date: Thu, 28 Aug 2008 09:40:11 +0100
Message-ID: <7765c8970808280140x3c7808d3kabd3f4ef40a5e96@mail.gmail.com>
Well 2 weeks ago is supiciously close to Patch Tuesday when updates for MS
products including Access were released. - There isn't anything about the
specific patch that would immediately seem to match (
http://www.microsoft.com/technet/security/bulletin/ms08-041.mspx) but still
it would seem a bit coincidental, especially if the patch were rolled up
with say a service pack to office.
Niall
On Wed, Aug 27, 2008 at 10:54 PM, Rumpi Gravenstein <rgravens_at_gmail.com>wrote:
> First here's what I'm running on:
>
> Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit
> PL/SQL Release 10.2.0.3.0 - Production
> CORE 10.2.0.3.0 Production
> TNS for Linux IA64: Version 10.2.0.3.0 - Production
> NLSRTL Version 10.2.0.3.0 - Production
>
> About a week or two ago a user started to complain that his Access 2002
> Oracle queries started slowing down, not by a little, but by a lot.
> Something that used to take a few minutes was now running into the hours.
> In this first case, the query has just one bind variable, a date, all the
> other predicate conditions are column joins or literals.
>
> So we ran a trace on it and got that the Access query was suddenly using a
> very bad explain plan.
>
>
>
> Everyone claims that nothing has changed but clearly something has as the
> Access run of the query now always runs slow and a cut and paste of the same
> query through Toad always runs fast.
>
>
>
> Also interesting, is that other users of Access databases that go against
> different tables have started to report slow-downs as well, but
> interestingly not all at the same time. I checked and am told that there
> have been no updates that should affect the Access 2002 to Oracle connection
> environment in the last few weeks. Having said that, about 1.5 months ago
> there was a migration to the 10g client.
>
>
>
> I'm going to run a 10053 trace against the Access version of the query to
> see if that sheds any light on the issue.
>
>
>
> In the mean time I thought I'd post here to see if anyone else has seen
> something like this.
>
>
>
> --
> Rumpi Gravenstein
>
-- Niall Litchfield Oracle DBA http://www.orawin.info -- http://www.freelists.org/webpage/oracle-lReceived on Thu Aug 28 2008 - 03:40:11 CDT