Re: Tuning Question

Daniel Morgan wrote:
> Burton Peltier wrote:
>
>> Good question... none :)
>>
>> Anyway, if you missed the original post, I have someone who works
>> with me who is asking to make the change. He is running some new
>> monitoring tool ( some free scripts) and it points at this 1 query
>> (and several others) as needing an index.
>>
>> --
>>
>> <Snipped>
>
> Then do absolutely nothing.

Why? Do you know why he is running it?

> More messes have been made fixing things that aren't broken than just
> about anything else.

So what prevents those guys from dropping the index if it doesn't work out?

Right now I'm in a comparable situation to this dba. I'm the only guy who actually sees performance data because everyone else is absolutely clueless. (Dunno whether the cluelessnes is comparable there, but here it's a fact. Just db users, no developers right now.)
So, my boss hears the complains but is unable to do anything about it and thinks it's the HW anyways, our users grumble quietly but have no idea if it has to be that way and the last thing I'd need is for some guy to go whining about whether he should follow the advice of his dba. Dba suggested it, the user should make clear that the dba assumes responsibility for that change and that's it.

As for that particular change, I had this too. We installed some software together with a db schema and it ran really slow. So, while a bigger server was being discussed I had a look at the schema and it turned out that one rather big table had no indexes at all. Looked at the problematic query, asked whether anyone had a problem with me tryin out a coupla indexes and voila - five minutes later performance was all right.

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"