Re: [Sqlrelay-discussion] Bug on php pear module - and testing

[David M. <dav...@fi...>]

On Wed, 23 Feb 2005 18:28:58 +0100
"vieonet forums" <df...@vi...> wrote:

> I have to change on line in php pear module, as it send error when
> $this->affectedRows == 0 if fact it's the following
> $this->affectedRows != 0
> 

What line is this on?  I can't seem to find a case where the module
returns an error when affectedrows == 0.


> Second I made test with http_load with a simple script using direct
> sqlrelay php function (no pear), The script is just injecting records
> in Mysql Heap memory table. I use the same insert for sqlrelay and for
> sqlb The result are the folowing : 
> http_load set on 1000 simultaneous users, during 30 seconds :
> 
>     sqlrelay has been able to insert as best parameter I could 921
>     lines sqlb has been able to insert 9020 lines
> 
> It's normal ?
> 
> Why so much differents ?
> 
> The second thinks is that on sqlb all the line are inserted when the
> test is end. Which is not the case on sqlrelay, the insert carry on
> after the test is over.
> 
> I do the testing on the same machine, with the same memory usage.

Some people have reported this problem recently and I made some
modifications to improve performance.  I'm going to make another
pre-release soon that will contain these modifications.

> 
> Thanks for any more information
> 
> david