[Koha] [Koha-devel] Performance of Koha and usage of remote MySQL server

[Fridolyn SOMERS]

Hie,

You have to consider it with the number of users connected at the same time.

For earch user connected :
MySQL server will require some CPU, a lot of RAM and some hard disk
performance.
Apache will need good CPU performance (to execute Perl).

So it depends on your hardware.
If you have simple dual-core systems with feew RAM, you should use a
dedicated server for MySQL.
It will be performant to serve many users, but with some waiting time.

If you have a big multi-core system with lot of RAM, you can use one server.
In this case, you may use a dedicated hard disk for MySQL, to have a good
read/write performance.

In addition, I think CPU frequency is important to have a fast
compilation+execution of Perl script.

In conclusion, I'd say that the problem should be cured at its source.
I mean that the number of SQL queries and cache systems should be optimized
in Koha code.
I think that database should always be on a dedicated server, it is nearly
an OS itself and needs all hardware resources.

PS : have a look at PCI Express 10Gb/s Ethernet cards.

Best regards,

On Fri, Apr 6, 2012 at 12:53 AM, Dobrica Pavlinusic <dpavlin at rot13.org>wrote:

> On Thu, Apr 05, 2012 at 02:03:02PM +0200, jonathan DUPERRIER wrote:
> > Hello,
> >
> > I would like to know if the installation of Koha with a MySQL
> > database on another server than the application may affect the
> > performance of Koha. The MySQL server and application are on the
> > same LAN, regardless of any order might be slowdown during query
> > execution.
>
> It seems it does.
>
> I did some measurements with DBD::Gofer which is network serialisation
> layer which can introduce local cache with:
>
> DBI_AUTOPROXY=dbi:Gofer:transport=null;cache=1
>
> This pass SQL query to original database, but cache queries in memory,
> just like MySQL query cache does.
>
> This has performance impact on generated search page:
>
> 2.952961 s page load time and 0.713 s of DBI time to
> 3.709371 s page load time and 1.634 s of DBI time
>
> This are plack numbers, but are useful for estimating performance
> implact of Gofer itself. It roughly doubles number of perl statements
> execuited in NYTProf.
>
> And now numbers from caching layer: we issue 1674 queries of which 927 are
> unique. That means we are generating double traffic to database so using
> local caching via DBD::Gofer (which is part of standard DBI install,
> BTW) might buy you some performance.
>
> There are some low-hanging fruit to pick here, I think :-)
>
>
> --
> Dobrica Pavlinusic               2share!2flame
> dpavlin at rot13.org
> Unix addict. Internet consultant.
> http://www.rot13.org/~dpavlin
> _______________________________________________
> Koha-devel mailing list
> Koha-devel at lists.koha-community.org
> http://lists.koha-community.org/cgi-bin/mailman/listinfo/koha-devel
> website : http://www.koha-community.org/
> git : http://git.koha-community.org/
> bugs : http://bugs.koha-community.org/
>



-- 
Fridolyn SOMERS
fridolyn.somers at gmail.com
Marsillargues - France
 <fridolyn.somers at gmail.com>