I already explained that, please reread my original post.

  _____  

From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Brian Kotek
Sent: Saturday, December 16, 2006 1:55 PM
To: [email protected]
Subject: Re: [Reactor for CF] URGENT... Survey app barfing under load


Did you load test it prior to release?


On 12/15/06, Dan Wilson <[EMAIL PROTECTED]> wrote: 

Marc,

Please post any and all frameworks you are using.

Also, run the Var Scope Checker tool on your Production Code:
http://www.petry-johnson.com/files/VarScopeChecker_v1.0.6.zip
<http://www.petry-johnson.com/files/VarScopeChecker_v1.0.6.zip> 

If you want some background, Dave Shuck said it better than I can:

http://daveshuck.instantspot.com/blog/index.cfm/2006/11/28/Thread-safety-exa
mple--Var-scope-your-loop-index-in-ColdFusion-CFCs


Unscoped variables can slam your code so lets start with the easy stuff
first.

Dan Wilson







On 12/15/06, Marc <[EMAIL PROTECTED]> wrote:
> Well, we got our survey app online, after finally figuring out that we 
> needed reactor to be in a directory in the root of our app with no
mapping.
>
> The client broadcast to his mailing list to take the survey, and lo and
> behold, our robust server is just choking... And it's really not THAT many

> users at once.  Of course, it's all my fault, so I need to ask for some
help
> at this point.  I didn't even think to ask him how many people were
getting
> the newsletter/notice of the survey, and reactor had been performing well
in 
> my tests -- though I did not have time to perform proper load tests on
this
> app, nor do I really have the tools to do so.
>
> We are in production mode.  There's a lot of questions in the survey, but 
> they are delivered in "pages".  The database server is not really under a
> lot of load overall, so this really seems to be a CPU usage issue.
>
> I'll answer any questions that anyone willing to help would like to ask,
to 
> help us get this completed app running better than it is.  I think we need
> to look at caching the queries that pull the questions onto each page as
the
> page is loaded, and we may need to look at how the answers are stored back

> to the database as well.  For each answer on the page, there's a database
> insert. I can also provide our dB schema.
>
> But first I thought I'd check with everyone, to see if there's some
> additional "performance related setting" I should check, within reactor. 
> Caching, mode, etc.
>
> I'm not sure if this should be taken off-list or not, but there will
> certainly be a reply back to the list when it's OVER, to make sure nobody
> makes the same mistake I have just made. 
>
> Anyone care to join me in hell's cubicle for a little while?
>
> marc
>
>
>
> -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
-- --
> Reactor for ColdFusion Mailing List 
> [email protected]
> Archives at: http://www.mail-archive.com/reactor%40doughughes.net/
> -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
-- --
>
>


--
"Come to the edge, he said. They said: We are afraid. Come to the
edge, he said. They came. He pushed them and they flew." 

Guillaume Apollinaire quotes


-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
-- --
Reactor for ColdFusion Mailing List
[email protected]  <mailto:[email protected]> 
Archives at: http://www.mail-archive.com/reactor%40doughughes.net/
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
-- -- 





-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
-- --
Reactor for ColdFusion Mailing List
[email protected]
Archives at: http://www.mail-archive.com/reactor%40doughughes.net/
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
-- --



-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Reactor for ColdFusion Mailing List
[email protected]
Archives at: http://www.mail-archive.com/reactor%40doughughes.net/
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

Reply via email to