Re: [Zope] Medusa performance - was: HELP: downloading file from netscape/old IE can frozezope
Hello, I have used medusa to build a 'host monitoring' system (very easy:I just had to code a module). And I find medusa extremly fast : I have tested it at ~3 millions hit/day(~33hit/sec) whithout problem, actualy I don't know what is the top limit on my config(dec alpha 433). BUT this was only with very small pages. What I dont understand with zope2 is why I have 7 process started: 'python z2.py -D' ?? If I understoud correctly medusa, it's running in a single process, so does it means that Zope use 1 medusa http sever + N (6??) 'zope process' to handle up to N (6?) requests simultanously? How does the Zope2/medusa part works? Gilles Lavaux
-----Original Message----- From: John Glossner <glossner@hotmail.com> To: zope@zope.org <zope@zope.org> Cc: Paul Everitt <Paul@digicool.com>; evan@4-am.com <evan@4-am.com> Date: Thursday, October 28, 1999 8:34 PM Subject: Re: [Zope] Medusa performance - was: HELP: downloading file from netscape/old IE can frozezope
--- Paul Everitt Paul@digicool.com wrote: --- Actually the strategy we are using with zope.org and Apache write now is slower (PCGI) than what we're migrating to starting today (mod_fastcgi).
Actually, I guess the real intent of the question was how many pages/sec can a Medusa server handle versus Apache? At what point do you decide to cut over? I ask this out of ignorance.I have never worked with a high volume site. I just found it somewhat odd that zope runs off of Apache as the entrance to the site.
Thanks, John
At 14:04 29/10/99 , Gilles lavaux wrote:
Hello,
I have used medusa to build a 'host monitoring' system (very easy:I just had to code a module). And I find medusa extremly fast : I have tested it at ~3 millions hit/day(~33hit/sec) whithout problem, actualy I don't know what is the top limit on my config(dec alpha 433). BUT this was only with very small pages.
What I dont understand with zope2 is why I have 7 process started: 'python z2.py -D' ?? If I understoud correctly medusa, it's running in a single process, so does it means that Zope use 1 medusa http sever + N (6??) 'zope process' to handle up to N (6?) requests simultanously?
How does the Zope2/medusa part works?
Gilles Lavaux
Zope is multi-threaded, and on certain platforms, threads are listed as seperate processes. IIRC, one thread for medusa, one thread that handles all other Zope threads, and N threads for handling requests, where N defaults to 4. Medusa answers the call, one of the Zope threads picks it up and the result is returned by Medusa. This way multiple requests can be taken care of simultaneously. -- Martijn Pieters, Web Developer | Antraciet http://www.antraciet.nl | Tel: +31-35-7502100 Fax: +31-35-7502111 | mailto:mj@antraciet.nl http://www.antraciet.nl/~mj | PGP: http://wwwkeys.nl.pgp.net:11371/pks/lookup?op=get&search=0xA8A32149 ------------------------------------------
Gilles lavaux wrote:
What I dont understand with zope2 is why I have 7 process started: 'python z2.py -D' ?? If I understoud correctly medusa, it's running in a single process, so does it means that Zope use 1 medusa http sever + N (6??) 'zope process' to handle up to N (6?) requests simultanously?
You are (presumably) running linux, which assigns a pid to each thread. Medusa is single-process multi-threaded. You can specify in your start file how many threads you want ( -t #threads ) <- I think. The default is 6, + the original process = 7. -- Ethan "mindlace" Fremen you cannot abdicate responsibility for your ideology.
participants (3)
-
Ethan Fremen -
Gilles lavaux -
Martijn Pieters