Re: postgresql.conf (Proposed settings)

On Thursday 22 November 2001 06:52, mlw wrote:

> > For this kind of "newbie", the kind that doesn't read the
> > documentation, this would only make it worse, because they'd assume
> > that by making the choice between three default configurations they've
> > done an adequate amount of tuning.  Basically, you'd exchange, "I
> > can't find any tuning information, but it's slow" for "I did all the
> > tuning and it's still slow".  Not a good choice.

I am the other kind of newbie, the one that reads documentation.
However, I fail to find much regarding tuning within the documentation 
delivered with the original version 7.1.2 tarball. Even when searching the 
postgresql website or the mailing list archives, the information is still 
sporadic. In the whole "Administrators guide" section there is no chapter 
"tuning", and even if you grep your way through it, you won't find much. 
And yes, I have read Bruce's book , from first to last page. And I have 
read Stinson's PostgreSQL Essential Reference the same way. Still, I am no 
wiser.

I am not complaining. Postgresql, after all, is free. But it is kinda 
strange to blame the users for not reading documentation if this 
documentation is that hard to find (does it exist?) that "Joe Average" 
can't find it.

The other domain where I could hardly find any information at all is the 
Postgres log settings.

I would be more than happy to write the documentation if I could get hold 
of the neccessary information, as I badly need it myself.

The other thing I am thinking abouty is that tuning can't be any "magic". 
If I can tune it, why shouldn't a configuration script be able to do the 
same? After all, even I would follow some algorithm to do it; it would 
probably involve a few educaterd trial & error & experiments - but hey, a 
script can do that too, can't it? I know too little about Postgres to do 
that part myself, but I don't think it is valid just to shove the idea of 
"autotuning" aside like this.

Horst