I would:
* Up the number of simultaneous requests.
* Check how much Ram does your server has.  Is it a dedicated CF server?
You can probably increase your heap size.  
* Fork over the 300 bucks for SeeFusion Enterprise (Not that much,
really)
*... and enable the database logging.  The counters table will provide
you the history of exactly how many running and queued requests you had
and the memory usage at the time of the error etc...

Also, a note-- you said "I have not gone above 8 simultanious requests".
I'm not sure how Fusion Reactor works, but queued requests may not be
part of that number.  In other words, you quite possibly are having more
than 8 requests at once, BUT only 8 of them are running at once because
the rest are queued.  (and the queued ones are the ones timing out.)

~Brad


-----Original Message-----
From: J W [mailto:[EMAIL PROTECTED] 
Sent: Monday, September 17, 2007 1:25 PM
To: CF-Talk
Subject: Re: Server hanging. Wierd problems. HELP!

I can't tell exactly what happened at the time of last crash to coincide
since the trial of SeeFusion I downloaded had a 2hr time limit on it..
However I just installed fusion reactor trial on the system

On this server

8 Max simultaneous requests
100mb max post data

JVM Arguments

512 Max Heap size (MB)
-server -Dsun.io.useCanonCaches=false -XX:MaxPermSize=128m -
Dcoldfusion.rootDir={application.home}/../ -
Dcoldfusion.libPath={application.home}/../lib -Xdebug
-Xrunjdwp:transport=dt_socket,server=y,address=28999,suspend=n

On the fusion reactor side:

I have a butload of slow request, but none are stuck. Alot of the slow
requests are waiting for communication with various 3rd party interfaces
to
complete, so the templates can take a long time to generate. Slow
requests
range from 4,000 to 35000ms. most are in the 4-5k range.

In the first 20 minutes of uptime here.

I have not gone above 8 simultanious requests, mostly 3-5.
Recent slow pages was up to 5.
Average Request time various between 1100 and 2800ms
Currently using 25mb of the heap. 4% memory for now.

23 minutes of uptime 1300 requests..

Where can we go from here?



On 9/17/07, Brad Wood <[EMAIL PROTECTED]> wrote:
>
> What was the memory heap size?  In other words, was it 70% of 150
Megs,
> or 70% of 1,500 megs?
>
> In ColdFusion administrator, what is the maximum number of
simultaneous
> requests set to?  (default is 15 to 20 I think).
>
> How many requests were running at the time the error was encountered?
>
> ~Brad
>
> -----Original Message-----
> From: J W [mailto:[EMAIL PROTECTED]
> Sent: Monday, September 17, 2007 12:37 PM
> To: CF-Talk
> Subject: Re: Server hanging. Wierd problems. HELP!
>
> I had Seefusion running. NOTHING looked out of the ordinary. I had
some
> slow
> processes, but that expected for some of the third party calls that I
> have
> to do on those. Memory was like 70=80%,  Is there anything specific I
> should
> be looking for using either SeeFusion or Fusion reactor?
>
>
> On 9/17/07, Brad Wood <[EMAIL PROTECTED]> wrote:
> >
> > I agree with using SeeFusion or Fusion Rector to see what is going
on.
> >
> > In the mean time, cfstat can at least show you if you are getting
> backup
> > up requests.
> >
> > ~Brad
> >
> > -----Original Message-----
> > From: John Mason [mailto:[EMAIL PROTECTED]
> > Sent: Monday, September 17, 2007 11:30 AM
> > To: CF-Talk
> > Subject: RE: Server hanging. Wierd problems. HELP!
> >
> > You need to use seefusion or fusionreactor to see what exactly is
> > happening.
> > Sounds like a hung cf page that is filling up your available CF
> threads.
> >
> >
> >
>
>
>
> 



~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~|
ColdFusion 8 - Build next generation apps
today, with easy PDF and Ajax features - download now
http://download.macromedia.com/pub/labs/coldfusion/cf8_beta_whatsnew_052907.pdf

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:288617
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Reply via email to