Re: Deadly OPTIMIZER_SECURE_VIEW_MERGING=TRUE

On Dec 5, 6:36 am, Noons <wizofo..._at_yahoo.com.au> wrote:
> On Dec 5, 8:00 am, Ben <benal..._at_yahoo.com> wrote:
>
> > I just spent about 20 minutes trying to sift through the blog, ( it is
> > hell trying to find anything on a blog unless you know the approximate
> > timeline ) and did a google search for cdos ( which just gets me debt
> > consolidation sites ). anyway I could get a link or unabreviated name?
>
> Sorry for that, Ben. Took me a while
> to come back here: been locked in courses
> the last two days.
>
> You can read about someone else's
> problems with it as well here:
>
> http://christianbilien.wordpress.com/2007/03/
>
> post of March 14th (just in case I get the usual
> interjection insinuating "somehow I'm the only one
> with this problem"...)
>
> The Metalink note mentioned there talks about it.
> The bug mentioned in someone else's post
> also explains what the problem is.
>
> I personally opened two calls with Oracle nearly
> three years ago with precisely the same problem,
> first with 9i, then with 10g. Which were closed by
> Oracle Support as "known problem, fixed in 11g".
>
> Boy, did this cause us a problem!...
> Try to explain to google for example that 3.5
> million of their clicks just didn't get accounted
> for because they got aggregated into another
> table. They don't like that sort of explanation...
>
> 3(three) years ago! And it was a "known" bug
> waaaaaay back then...
> In fact, it's been "known" since the auto-sga
> started, in 9i early releases.
>
> It is absolutely ludicrous it hasn't yet been fixed
> in ALL available releases. Not just 11g
> (*IF* it is fixed there...).
>
> And then some folks have the gall to call
> me cynical when I point out that this is
> inadmissible in a product that:
>
> 1- is charged for, not freeware
> 2- that claims to be "enterprise class"
> 3- that claims to be "unbreakable"
> 4- that is pushed as multi-schema,
> multi-aplication capable.
>
> As a matter of additional info and to end in
> a positive note: read the whole of Christian's
> blog, it is well worth someone's time no matter
> how much they know about Oracle.
>
> Kevin Closson mentioned it a while
> ago and I thank God he did!

thanks to all for pointing me to the posts. That is amazing to me. We run JDE and there are probably 3 data schemas that have identical tables, 3 control data schemas with identical tables, and 3 more metatype  data schemas with identical tables. Yet, I've yet to hit the problem. I'd would have to say it is probably because the ERP system uses fully qualified sql statments.

Do you know if this can happen if you are logged in as one of the schema owners? Is this limited to a certain platform? Received on Wed Dec 05 2007 - 12:08:31 CST