Re: Scaling for performance : Results and comments????

[Andreas <andreas@conectiva.com.br>]

On Fri, Jun 11, 2004 at 11:47:11PM -0700, Trevor Warren wrote:
>   Why not. I can let you know the methodology and how
> these were conducted  but these tests require tools
> like Load Runner and QA Load which run into a couple
> of 100K$ is you can afford em.

Ugh :)

>   We are a benchmark lab here and due to our good
> relationship with these vendors we get the best
> testing tools the industry can afford.

I see

>   I think it would be better for me to share the
> methodology with you and instead of using the scripts
> of these expensive tools i can help educate you on
> using Open Source tools towards benchmarking the
> System under various configuations. 

I'm not sure I will be able to reproduce your tests then.
After all, if I could, those 100K$ tools would be a blatant
ripoff then :)

>   Lemme know and we shall proceed. But inturn for all
> my feeding you info i would request the results of
> your tests to be published here on the openldap
> mailing list please. If its okie with you lets
> proceed.

Of course, I already did this in the past and published here
my results. At that time, I was testing the effect of a well
tuned DB_CONFIG file regarding bulk loading (with indexes) 
of 1 million entries (standard RFC2307 schema used to represent
user accounts). This was with the 2.1.x series, I didn't play
with 2.2.x yet.