--
"Volker Hetzer" <volker.hetzer_at_ieee.org> wrote in message
news:bdafhr$7k0$1_at_dackel.pdb.sbs.de...
> Daniel Morgan wrote:
> > Your situation is not comparable to the OP's. You had user complaints
> > the OP had nothing but the recommendation of
> > some scripts of unkown (at least to us here) value. In your case you
> > had something specific to fix and a measure to
> > determine success or failure. All the OP can do is satisfy a tool or
> > create real problems.
>
> Personally I see much too few complaints. It's just my opinion but
> to wait til the users come to you is IMHO not a real good idea.
> Since the DBA sees much more of the internal oracle stuff he should
> go the the app developers and make suggestions.
> Waiting for users/developers assumes that they know the optimum behavior.
> And especially when it comes to speed and response time of the application
> I've yet to see the user who wouldn't want a faster app.
>
> But maybe we are talking about different environments here. If I were to
> administrate a db a nuke powerplant depends on I'd be very careful
> to touch a running system too.
>
> I, and probably you too know not enough about Burtons environment.
> Is the system loaded so much that for instance the impact of another
> index on DML time causes a degradation serious enough to forbid
> a test? Does he have a test environment with realistic load at all?
> Would giving a shout to the office and just doing it if no one answers be
> a problem?
> Is his app still in development? How much costs an hours downtime or
> data loss?
As I stated above, this is just something we will have to test in our test
environment, which is available...
Actually, after posting my original post, I can now see this as 2 issues:
1) Another DBA suggesting changing production without testing - I don't do
that for anything ... almost :)
2) Priorities and where do you draw the line in benefits for time spent on
this vs. time spent on other things .
>
> Lots of Greetings!
> Volker
> --
> While it is a known fact that programmers
> never make mistakes, it is still a good idea
> to humor the users by checking for errors at
> critical points in your program.
> -Robert D. Schneider, "Optimizing INFORMIX
> Applications"
>
>
>
Received on Wed Jun 25 2003 - 22:32:31 CDT
