cherrypy server stress tests

Hello CherryPy users,

I recently started working at a new company where we gonna
develop a
webapplication with high security and large amounts of
users.
I've already looked into the security issues and it seems
that, that
isn't really gonna be a problem to solve however i can't
really find
anything about load / stress testing of the cherrypy
server.

In order to compare cherrypy with the other options, I was
looking for
some type of document where connecting time, receiving and
sending
times are specified.

offcourse it's possible to perform these tests ourselves but
it would
be a waste of time if such information already exists. Does
any know
if such information is already available and where it's
located.

thanks in advance for your help,

Richard Mendes


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at h
ttp://groups.google.com/group/cherrypy-users?hl=en
-~----------~----~----~----~------~----~------~--~---

Hi

Check out:

http://www
.cherrypy.org/wiki/CherryPySpeed

Since I've started using CherryPy with mod_wsgi I've
found that CherryPy pretty much irrelevant in terms of
performance of my web app - it's the database that's
the bottleneck. I suspect that this will be the case
with virtually any web app.

A

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at h
ttp://groups.google.com/group/cherrypy-users?hl=en
-~----------~----~----~----~------~----~------~--~---

thanks,

Now i was able to show something of how the server would
handle
himself during heavy traffic.
I did some tests with 'ab' from apache as well and the load
handling
was alright althougt the server
errored when more than a hundred users simaltaniously
connected and
tried to get data.

Richard



On Oct 9, 9:53 am, Andrew Veitch <andrew.vei...yahoo.co.uk> wrote:
> Hi
>
> Check out:
>
> http://www
.cherrypy.org/wiki/CherryPySpeed
>
> Since I've started using CherryPy with mod_wsgi I've
> found that CherryPy pretty much irrelevant in terms of
> performance of my web app - it's the database that's
> the bottleneck. I suspect that this will be the case
> with virtually any web app.
>
> A


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at h
ttp://groups.google.com/group/cherrypy-users?hl=en
-~----------~----~----~----~------~----~------~--~---

On 10/9/07, cyclopsvs <mendes150gmail.com> wrote:
> Now i was able to show something of how the server
would handle
> himself during heavy traffic.
> I did some tests with 'ab' from apache as well and the
load handling
> was alright althougt the server
> errored when more than a hundred users simaltaniously
connected and
> tried to get data.

If you haven't already, you should try tuning the server by
playing
with the following config settings:

server.socket_queue
--------------------------
This tells the app server the number of connections to queue
while the
worker threads are busy handling other requests.  After the
socket
queue is full, any further connections will be denied.  The
default is
5, IIRC.  See the following post for some eye-opening
results when
tweaking server.socket_queue.
http://www.aminus.org/blo
gs/index.php/fumanchu/2007/10/02/cherrypy_3_request_queue_si
ze

server.thread_pool
------------------------
The number of worker threads available to handle requests. 
The
default is 10.  More threads will allow more simultaneous
requests to
be handled, albeit with greater resource consumption.  But
unused
resources are wasted resources, right? 

Also note that the CherryPySpeed wiki page is quite out of
date.  It
looks like it was written for CherryPy 2.0.  Maybe someday I
(or
someone else) will get a chance to update it.

In the meantime, take a look at
http://www.aminu
s.org/blogs/index.php/fumanchu/2006/12/23/cherrypy_3_has_fas
test_wsgi_server_yet
That is nearly a year old, but it at least has some numbers
on
CherryPy 3's WSGI server.  There is a host of other good
info on
Robert's blog.  Have a look around there.

HTH,

Christian
http://www.dowski.com

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at h
ttp://groups.google.com/group/cherrypy-users?hl=en
-~----------~----~----~----~------~----~------~--~---

Thank for these suggestions,

Seems like there can be done a lot of tuning to create a
optimal
server condition. I will read op on all these suggestions
and start
tuning the server config.

By looking at this it seems that cherrypy might be the best
option to
use in this case, see if this will convince the system
architect and
project manager.

thanks.

Richard

On 10/9/07, Christian Wyglendowski <christiandowski.com> wrote:
>
> On 10/9/07, cyclopsvs <mendes150gmail.com> wrote:
> > Now i was able to show something of how the server
would handle
> > himself during heavy traffic.
> > I did some tests with 'ab' from apache as well and
the load handling
> > was alright althougt the server
> > errored when more than a hundred users
simaltaniously connected and
> > tried to get data.
>
> If you haven't already, you should try tuning the
server by playing
> with the following config settings:
>
> server.socket_queue
> --------------------------
> This tells the app server the number of connections to
queue while the
> worker threads are busy handling other requests.  After
the socket
> queue is full, any further connections will be denied. 
The default is
> 5, IIRC.  See the following post for some eye-opening
results when
> tweaking server.socket_queue.
> http://www.aminus.org/blo
gs/index.php/fumanchu/2007/10/02/cherrypy_3_request_queue_si
ze
>
> server.thread_pool
> ------------------------
> The number of worker threads available to handle
requests.  The
> default is 10.  More threads will allow more
simultaneous requests to
> be handled, albeit with greater resource consumption. 
But unused
> resources are wasted resources, right? 
>
> Also note that the CherryPySpeed wiki page is quite out
of date.  It
> looks like it was written for CherryPy 2.0.  Maybe
someday I (or
> someone else) will get a chance to update it.
>
> In the meantime, take a look at
> http://www.aminu
s.org/blogs/index.php/fumanchu/2006/12/23/cherrypy_3_has_fas
test_wsgi_server_yet
> That is nearly a year old, but it at least has some
numbers on
> CherryPy 3's WSGI server.  There is a host of other
good info on
> Robert's blog.  Have a look around there.
>
> HTH,
>
> Christian
> http://www.dowski.com
>
> >
>

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the
Google Groups "cherrypy-users" group.
To post to this group, send email to cherrypy-usersgooglegroups.com
To unsubscribe from this group, send email to
cherrypy-users-unsubscribegooglegroups.com
For more options, visit this group at h
ttp://groups.google.com/group/cherrypy-users?hl=en
-~----------~----~----~----~------~----~------~--~---