Re: Multiple databases - best performance scenario

From: joel garry <joel-garry_at_home.com>
Date: Wed, 16 Apr 2008 10:59:04 -0700 (PDT)
Message-ID: <ce403f47-ab5e-473f-a2d0-4de3a3c95ebe@p25g2000pri.googlegroups.com>


On Apr 16, 5:38 am, bhonaker <bhona..._at_gmail.com> wrote:
> On Apr 15, 1:02 pm, NetComrade <netcomradeNS..._at_bookexchange.net>
> wrote:
>
> > On Tue, 15 Apr 2008 06:48:03 -0700 (PDT), bhonaker
>
> > <bhona..._at_gmail.com> wrote:
> > >Wow what a vicious bug.  This bug does not appear when using fully
> > >qualified objects, which should be the goal anyway :-).
>
> > Be realistic. And as far as bugs go, it was very embarassing when I
> > told developers "common, that's what database is build for, for this
> > to never happen", only to be shown proof that I was a fool thinking
> > that.
> > .......
> > We run Oracle 9iR2,10gR2, 10g2RAC on RH4/RH5 and Solaris 10 (Sparc)
> > We use RMAN and remote catalog for backups
>
> I'm not saying its not difficult, but unrealistic?  When compiling
> code into production, use a DBA account, not a schema that owns
> production code - unqualified objects shake themselves out.  At least
> that's how we've done it at my last couple gigs.

I haven't thought this totally through, but somehow it strikes me as giving objects more authority than they need. But then again, whenever I do this at a new place I think it totally through, because I always forget something, and there's always some newfangled rule. And of course, I do it totally wrong now thanks to app vendor strangeness. And I'm not talking quarks.

jg

--
@home.com is bogus.
Must be the lack of water.  http://www.signonsandiego.com/uniontrib/20080416/news_1n16beertour.html
Received on Wed Apr 16 2008 - 12:59:04 CDT

Original text of this message