About MC/RR applications servers

Pierre Sahores psahores at easynet.fr
Sat Jul 12 20:12:00 EDT 2003


On Sun, 2003-07-13 at 01:26, revolution at knowledgeworks.plus.com wrote:
> Thanks Pierre
> 
> >>
> Just do tests and you will see as me that this works perfectly, faster
> for example, than in using ASP's or PHP commands, because the linux bash
> optimisations, because the psql perfect design to work in command-line
> pipe mode.
> <<
> Well, this is rather flabberghasting...
> I suppose it does fit in with the unix philosophy.  I am prepared to believe you, and when I get the time I will try to do some comparisons for myself.
> 
> >>
> I just remplace all of them by metatalk/transcript calculations on very
> simple SQL requests replies
> <<
> With regard to SPs, triggers, etc.. I still don't understand why you do not use them.  Is it because they do not work, because they are slow, or because you just prefer to do it in Transcript?

1.- In avoiding to use them, even if they can works, i am able to handle
a best "abstract layer" between the database back-end and the
applications server. This is always very usefull when we don't know, at
the beginning of the discussions with the client, what the application
will look like exactly, to the end. It's why, i will always be more
confortable in using metatalk/transcript handlers instead.

> >>
> 75% of writes on databases going from some megs (events statistics) up
> to 250 Megs of datas, with, as an average, no more than 10 Megs peer
> table
> <<
> And it is basically a write-intensive application?  
> Do I understand correctly that each write can be several megabytes?

No. individual writes POSTS don't never contains more than about 100 to
150 ko each, 

> And that the database size is around 250mb?

250 Mb peer year, yes.
> 
> >>
> Suse 8 Pro x86 - Athlon 800 1 Go Ram - server. All the other
> apps are hosted on 3 st box, again Suse 8 Pro x86 - Athlon 800 1 Go Ram.
> <<
> Each server has 1gb of RAM?  Do you find you need that much?
> I suppose if your queries are write-intensive this extra RAM is not there to boost PostgreSQL's cache?

With 512 Mo, it would be enough but i tuned the servers to 1 Go when
there was doubts about the memory usage of the the earlies issues of the
linux 2.4.x kernel.
> 
> I'm in agreement about OS X being slow compared to Linux.  Imagine running Linux on Panther :-)

As an example, the "sockettimeoutinterval" global needs to be set to
30000 under Jaguar, to have correct "pg_dump" and "restore.sql"
processing. Under Linux, the same global is just set at his default
startup value.
> 
> Bernard  
> 
> _______________________________________________
> use-revolution mailing list
> use-revolution at lists.runrev.com
> http://lists.runrev.com/mailman/listinfo/use-revolution
-- 
Bien cordialement, Pierre Sahores

Serveurs d'applications & bases ACID SQL
Penser et produire l'avantage compétitif




More information about the use-livecode mailing list