Re: tkprof question

Hi Norman,

Ta muchly.

I do remember reading a thread semi-recently regarding this. It all sounds rather yucky poo.

And they saying that "touching wood" is superstitious nonsense ;)

Cheers

Richard
"Norman Dunbar" <Norman.Dunbar_at_lfs.co.uk> wrote in message news:E2F6A70FE45242488C865C3BC1245DA702FAE411_at_lnewton.leeds.lfs.co.uk...
> Hi Richard,
>
> timed_stats normally doesn't make any difference to my set up and I
> always run with it on. However, recently a customer of our sister
> company was having problems and I asked for half our statspack reports
> with times stats on - the whole bleeding thing fell over !
>
> Responses were abysmal, stuff was timning out etc and users complained
> like b*gg*ry !
>
> I got enough evidence to point squarely at the discs as being the
> bottleneck - which it turned out was indeed the case. (Hooray for me!)
> but setting T_S has a very serious impact on the system.
>
> Jonathan pointed me in the direction of a bug in Oracle whereby under
> certain circumstances, having T_S turned on, caused sharable SQL to be
> not shared - and I think this was the problem we eventually hit as well.
> I've asked them to patch to 8174, but as they are a customer of our
> sister company, internal politics must reign, and they have said not to
> patch yet - oh hum !
>
> Cheers,
> Norman.
>
> PS. It was a Sun server of some kind - I can't remember exactly which
> one and Oracle 8172 as well.
>
>
> -------------------------------------
> Norman Dunbar
> Database/Unix administrator
> Lynx Financial Systems Ltd.
> mailto:Norman.Dunbar_at_LFS.co.uk
> Tel: 0113 289 6265
> Fax: 0113 289 3146
> URL: http://www.Lynx-FS.com
> -------------------------------------
>
>
> -----Original Message-----
> From: Richard Foote [mailto:richard.foote_at_bigpond.com]
> Posted At: Wednesday, November 27, 2002 12:02 PM
> Posted To: server
> Conversation: tkprof question
> Subject: Re: tkprof question
>
>
>
> Hi Andy,
>
> I yet to see timed_statistics have any discernable impact on
> performance. Do
> you have any particular environment or configuration where you've
> measured
> anything as significant as 2-3% ?
>
> Cheers
>
> Richard
>
>
Received on Wed Nov 27 2002 - 06:47:05 CST