Re: JRun errors after security patch install

[Terry Ford]

That's the problem -- thanks for the response!

"JRun too busy or out of memory" as an error message for "form variable limit 
reached" could probably be improved upon ;)


Regards,
Terry




 From: Carl Von Stetten 
To: cf-talk  
Sent: Wednesday, January 16, 2013 8:49 AM
Subject: Re: JRun errors after security patch install
 

Terry,

I assume you are referring to the latest patch released earlier this 
week.  Since it was a cumulative patch, it included some previously 
released security improvements as well as the new ones.  One of the 
older security improvements was to add a "Post Parameter Limit" setting 
to prevent denial of service (DOS) attacks using hash algorithm 
collisions.  You can find details of the issue here 
<http://www.adobe.com/support/security/bulletins/apsb12-06.html>. The 
default value for this new setting is 100, but you can increase it if 
you need to.  Instructions to do this can be found here 
<http://helpx.adobe.com/coldfusion/kb/coldfusion-security-hotfix.html>.

HTH,
-Carl V.
On 1/16/2013 7:36 AM, Terry Ford wrote:
> Hello --
>
> I installed the security patch last night on cf 9.1 linux, and woke up this 
> morning to find that some of our templates no longer work:
>
>
> [Wed Jan 16 09:28:27 2013] [notice] jrApache[1978: 53193]  returning error 
> page for JRun too busy or out of memory
>
>
>
> They return a 500 error to the user.
>
> The commonality in the templates that are doing this is that they have lots 
> of FORM variables.   So, they're dealing with a lot of form input.  They have 
> worked perfectly for years, up til this patch.
>
>
> Has anyone else experienced this?   We are running on linux/apache.
>
> Regards
>
> 



~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:353938
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm

JRun errors after security patch install

[Terry Ford]

Hello --

I installed the security patch last night on cf 9.1 linux, and woke up this 
morning to find that some of our templates no longer work:


[Wed Jan 16 09:28:27 2013] [notice] jrApache[1978: 53193]  returning error page 
for JRun too busy or out of memory



They return a 500 error to the user.

The commonality in the templates that are doing this is that they have lots of 
FORM variables.   So, they're dealing with a lot of form input.  They have 
worked perfectly for years, up til this patch.


Has anyone else experienced this?   We are running on linux/apache.

Regards

~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:353929
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm

Re: Getting basic CF8 CFCACHE working in CF9

[Terry Ford]

Thanks for the reply Ray.

Other attributes seemingly do nothing too, or don't work as expected.   

I've discovered that page-caching (CF8's default behavior) is possible through 
ehcache, but there's a surprising lack of information on how to implement it to 
mimic CF8.  CFCACHE is the first feature in 9 generations of CF that has me 
scratching my head why code was broken so badly from one version to another, 
with seemingly no documentation to lay out precisely how to achieve 
like-behavior. 

Also, I have discovered a way to effectively crash CF9 by using a malformed 
CFCACHE attribute.   Using the old, CF7 "timeout" attribute gives Jrun a heart 
attack, and effectively kills the server.  At least on linux / apache.

Add this code:
 

Then run the template a bunch of times.  On linux/apache at least the server 
becomes unstable and half the requests get Jrun errors until reboot.The 
"timeout" param was deprecated in CF7, but we still had some floating around in 
our code since CF8 played nice with it.   It took me hours to trace down the 
problem to CFCACHE.



> On Fri, May 13, 2011 at 10:54 PM, Terry Ford  
> wrote:
> > 1. I can't for the life of me work out how to do that in CF9.  First, 
> I need to add "usequerystring=true" right because the default behavior 
> was changed in CF9?
> 
> Right - I believe the idea is that before it was automatic, but in 9
> we allow you to disable that if you want.
> 
> > 2. Then, do I need to wrap the entire page with  ... 
>  ?   Or, is just a lone  statement at the top of 
> the page like in CF8 sufficient?
> 
> You can do either. If you just put it on top, it means cache
> everything. If you wrap, it means cache a fragment.
> 
> 
> > 3. Third, and most frustratingly, how do you CFCACHE to disk?  I'm 
> using this statement in CF9:
> >
> 
> I see the same. This looks like a bug unfortunately. Please log a
> report. If you absolutely want file based caching, you could do it
> manually of course. It could be coded up in about 30 minutes. My old
> custom tag, scopeCache (http://scopecache.riaforge.org/) would allow
> you to do 


~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:344521
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm

Getting basic CF8 CFCACHE working in CF9

[Terry Ford]

Hey...

I am trying to get a CF8 app working on CF9, and it was all seamless until I 
ran into CFCACHE.

I'm having problems getting CFCACHE to save anything to disk.  I understand the 
"default" behavior is now in-memory caching, but for now I'm just trying to get 
this thing to work as it did in CF8:  to disk.  

This was the code in CF8:



It simply saves the page to disk, one page per URL combination.  

1. I can't for the life of me work out how to do that in CF9.  First, I need to 
add "usequerystring=true" right because the default behavior was changed in CF9?

2. Then, do I need to wrap the entire page with  ...  ?   
Or, is just a lone  statement at the top of the page like in CF8 
sufficient?

3. Third, and most frustratingly, how do you CFCACHE to disk?  I'm using this 
statement in CF9:



However, when I run the page, it's caching to memory.  No files are appearing 
in "directory".

Please, what am I missing?   How do I get it to save everything to "directory" 
so that the cache will persist across server restarts?



~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:344508
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm

Upgrading CF8 to CF9 breaking existing lucene stuff?

[Terry Ford]

Hi, I'm having an issue in a port from CF8 to CF9.

We created our own simple lucene indexing stuff on CF8, and are simply trying 
to get it to work on CF9.

Right now I'm having problems with this line, which works fine under CF8:

luceneQuery = queryParser.parse(keyword, "body", analyzer);

It is returning error "The parse method was not found. Either there are no 
methods with the specified method name and argument types or the parse method 
is overloaded with argument types that ColdFusion cannot decipher reliably. 
ColdFusion found 0 methods that match the provided arguments. If this is a Java 
object and you verified that the method exists, use the javacast function to 
reduce ambiguity. "

If I do a CFDUMP on queryParser, there indeed is a parse() function listed.

Does anyone have any idea why CF9 is throwing the error above, and what we can 
do to fix it?  I'd rather not mess around with Solr for now -- just want to get 
our existing stuff working.

Thanks 

~|
Order the Adobe Coldfusion Anthology now!
http://www.amazon.com/Adobe-Coldfusion-Anthology/dp/1430272155/?tag=houseoffusion
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:344478
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm

Re: Daylight Savings & Scheduled Tasks

[Terry Ford]

>I usually use the following JVM arg (I'm in the central time zone):
>-Duser.timezone=America/Chicago

Hmm, thanks...  but am still a little confused.

What determines whether coldfusion/CFScheduler switches to daylight savings 
time?  
(A) the timezone settings in the OS or (B) settings in the JVM?

I have told the OS to ignore daylight savings time ("use GMT-6 all year, every 
year").  Is the JVM going to override that based on its own internal timezone 
data and say "hey, GMT-6 sounds like it's central time, so therefore it's time 
for daylight savings time", or will it respect the OS's timezone data and keep 
ColdFusion on GMT-6?

Thanks


~|
Want to reach the ColdFusion community with something they want? Let them know 
on the House of Fusion mailing lists
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:331725
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm

Re: Daylight Savings & Scheduled Tasks

[Terry Ford]

>I personally would not have DST switched on for a server, and have it
>disabled

Thanks for the response.   I just switched DST off on the server (RHEL linux: 
ln -s /usr/share/zoneinfo/Etc/GMT+6 localtime).

However, the scheduler.log is still picking up CST somehow after a CF reboot:

"task -- Rescheduling for :Thu Mar 11 22:45:00 CST 2010"

Do you have any idea how CF picks up server timezone?  I would have thought 
that the JVM would have picked up that change.  I'm not sure where CF is 
getting that CST from. 



~|
Want to reach the ColdFusion community with something they want? Let them know 
on the House of Fusion mailing lists
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:331630
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm

Daylight Savings & Scheduled Tasks

[Terry Ford]

I have had the following problem the last couple of time changes, and would 
like to know if there is a simple solution to fix it:

I have a number of tasks that I need to run just after midnight, local time 
(CT), whether it's standard CST or daylight CDT.  12:01 AM.   

I use CF scheduler with settings "run daily at 12:01 AM".   However, whenever 
we hit a daylight savings and switch between CST and CDT, ColdFusion tries to 
be smart and adjusts that starting time by 1 hour, which means that the tasks 
start running at 11:01 PM CT. 

What I'd like to say is "run daily at 12:01 AM CT", so that it runs at 12:01 
regardless of whether we're in daylight savings or not.   Is there a way to do 
this in scheduler?

The work around is to wait for daylight savings to occur and then go in and 
resubmit all of those tasks to again run at "12:01 AM".  It then starts working 
correctly until the next time change.  This, however, is a very annoying 
solution.

Thanks





~|
Want to reach the ColdFusion community with something they want? Let them know 
on the House of Fusion mailing lists
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:331625
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm

CFDIVs within custom tag problem

[Terry Ford]

Hi devs!

I have run into a strange problem that probably has a simple fix but I can't 
work it out.

I have tag cf_commentform  that displays a feedback cfform within a cfdiv.  The 
user enters comments, and the form submits inline.  Works great.

Some pages call cf_commentform more than once.   For example, a blog site where 
there are three entries on a page may have 3 of these called.  Each one has 
unique IDs.

However, when I view the source, I am getting the javascript includes loading 
multiple times on the page:

script type="text/javascript" src="/CFIDE/scripts/ajax/package/cfajax.js" 
/script
script type="text/javascript" src="/CFIDE/scripts/cfform.js"  /script
script type="text/javascript" src="/CFIDE/scripts/masks.js" /script
script type="text/javascript" src="/CFIDE/scripts/cfformhistory.js" /script
script type="text/javascript" _cf_loadingtexthtml=" div align='center' img 
src='/CFIDE/scripts/ajax/resources/cf/images/loading.gif'/";

The above appears THREE times, and the browser loads it 3 times.  This is 
clearly unnecessary.   

When I simply use the commentform code in a cfinclude this does not happen -- 
it's only when I use the cfdiv/cfform within a custom tag.

Does anyone have any ideas on how I can keep using a custom tag like this 
without all of these multiple javascript loads appearing on the calling page?

Thanks


~|
Want to reach the ColdFusion community with something they want? Let them know 
on the House of Fusion mailing lists
Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:331281
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/groups/cf-talk/unsubscribe.cfm

Re: CFCHART / x-axis format

[Terry Ford]

>Can you show a bit more code, specifically how you pass the XML to cfchart?

Hi Ray,






Query Q is (score,cnt), int columns.

and silver.xml (in charting/styles) :











Resultant chart example:  Am trying to get rid of the .0's:

http://www.funtrivia.com/trivia-quiz/Celebrities/Musicians-From-Louisiana-297040.html

Thanks 

~|
Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to 
date
Get the Free Trial
http://ad.doubleclick.net/clk;207172674;29440083;f

Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:319357
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: CFCHART / x-axis format

[Terry Ford]

>Can you show a bit more code, specifically how you pass the XML to cfchart?

Hi Ray,






Query Q is (score,cnt), int columns.

and silver.xml (in charting/styles) :











Resultant chart example:  Am trying to get rid of the .0's:

http://www.funtrivia.com/trivia-quiz/Celebrities/Musicians-From-Louisiana-297040.html

Thanks 

~|
Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to 
date
Get the Free Trial
http://ad.doubleclick.net/clk;207172674;29440083;f

Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:319356
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

CFCHART / x-axis format

[Terry Ford]

Hey folks,

CF8.  I am plotting on X-axis "scores" (values 1 to 10), and on the Y-axis the 
number of people scoring that score.  

However, CFCHART is formatting the "scores" as decimals:  1.0, 2.0,...,10.0.

I have tried all sorts of things in the associated xml stylesheet:







I *thought* that labelFormat would solve the problem, but it seems to be 
getting ignored.  No matter what I set labelFormat to, I still get 1.0,...

The stylesheet's titlestyle font is also being ignored.  No matter what I set 
the font to, it doesn't change.  However, if I change the orientation tag, the 
orientation changes, so I most certainly am using the right stylesheet. 

Does anyone have any experience with modifying the formatting and/or font of 
the X-axis?  At a loss...

Thanks 

~|
Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to 
date
Get the Free Trial
http://ad.doubleclick.net/clk;207172674;29440083;f

Archive: 
http://www.houseoffusion.com/groups/cf-talk/message.cfm/messageid:319353
Subscription: http://www.houseoffusion.com/groups/cf-talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: SQL injection attack on House of Fusion

[Terry Ford]

Nimda did not use SQL injection as any sort of primary vector.

SQL injection attacks have been around forever, but botnet/worm SQL injection 
attacks have really taken off pretty recently.  It has gotten so bad that even 
Microsoft recently released a security advisory article that doesn't deal with 
a major flaw in its own software.  That's rare ;)

http://www.microsoft.com/technet/security/advisory/954462.mspx



--- On Sat, 8/9/08, Jochem van Dieten <[EMAIL PROTECTED]> wrote:

> From: Jochem van Dieten <[EMAIL PROTECTED]>
> Subject: Re: SQL injection attack on House of Fusion
> To: "CF-Talk" 
> Date: Saturday, August 9, 2008, 6:24 AM
> Terry Ford wrote:
> > Pretty ingenious really, infecting websites via
> injection attack in order to infect clients with browser
> vulnerabilities.
> 
> In 2001 it was:
> http://www.cert.org/advisories/CA-2001-26.html Now it is 
> just business as usual.
> 
> Jochem
> 
> 

~|
Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to 
date
Get the Free Trial
http://ad.doubleclick.net/clk;203748912;27390454;j

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:310615
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: SQL injection attack on House of Fusion

[Terry Ford]

Hi Ray,

I used this rewrite in the root directory's .htaccess file (linux/apache)... 
try Windows equiv?  I've run into the same problem that you describe using 
global rewrites in years past, but this directory-based method seems to work 
reliably.


order allow,deny
allow from all

RewriteEngine On
. rewrites here 


Regards


--- On Sat, 8/9/08, Raymond Camden <[EMAIL PROTECTED]> wrote:

> From: Raymond Camden <[EMAIL PROTECTED]>
> Subject: Re: SQL injection attack on House of Fusion
> To: "CF-Talk" 
> Date: Saturday, August 9, 2008, 12:12 AM
> Hmm. I'm having no luck with this. I'm trying it on
> a blogcfc site, so
> it's being added after /, so I also added path_info
> 
> RewriteEngine on  
> RewriteCond %{QUERY_STRING} .*DECLARE.* [NC]
> RewriteRule ^(.*)$  /violation.htm
> RewriteCond %{PATH_INFO} .*DECLARE.* [NC]
> RewriteRule ^(.*)$  /violation.htm
> 
> And it's not picking up when it sees declare in the
> url.
> 
> On Fri, Aug 8, 2008 at 11:07 PM, denstar
> <[EMAIL PROTECTED]> wrote:
> > You can keep it in a different .conf file, and use the
> Include
> > directive, to pull it in wherever you need it, BTW.
> >
> > --
> > Employ your time in improving yourself by other
> men's writings, so
> > that you shall gain easily what others have labored
> hard for.
> > Socrates
> >
> > On Fri, Aug 8, 2008 at 10:05 PM, denstar
> <[EMAIL PROTECTED]> wrote:
> >> A simple look at the docs would state why, but it
> doesn't appear to
> >> work if you've got it "floating"
> (rewrites in general).  I don't
> >> remember, off hand.
> >>
> >> I'm pretty sure it will work in a Directory or
> Location block too, tho.
> >
> > 
> 
> 

~|
Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to 
date
Get the Free Trial
http://ad.doubleclick.net/clk;203748912;27390454;j

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:310597
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: SQL injection attack on House of Fusion

[Terry Ford]

<< Also, I'd like to explore the possibility of reporting compromised people to 
their ISPs. >>

Well good luck ;)

As I posted earlier, we have been hit by over 200,000 attack attempts over the 
past 2 days.

Here's the analysis from our last 195,264 attack attempts:

Our attacks over the past *24 hours* have originated from *12,007* different IP 
addresses.  Twelve THOUSAND.  That is not a typo.   This is an extremely large 
botnet, pure and simple.   These IP addresses appaer to be largely random folks 
who are using browsers with vulnerabilities.   

Each client, on average, makes 2-4 attack requests.

Here are the origin IPs with the most attacks:

| ip  | attacks | tmp1 | tmp2 |
+-+---+--+--+
| 203.160.1.40|  1246 | NULL | NULL |
| 203.160.1.70|   596 | NULL | NULL |
| 61.164.132.230  |   478 | NULL | NULL |
| 211.72.233.9|   471 | NULL | NULL |
| 203.162.3.159   |   462 | NULL | NULL |
| 211.72.233.8|   452 | NULL | NULL |
| 211.72.233.10   |   429 | NULL | NULL |
| 221.253.217.138 |   319 | NULL | NULL |
| 210.112.177.244 |   252 | NULL | NULL |
| 59.15.212.125   |   252 | NULL | NULL |
| 70.88.218.70|   240 | NULL | NULL |
| 67.86.134.184   |   234 | NULL | NULL |
| 125.107.109.47  |   231 | NULL | NULL |
| 202.92.190.172  |   225 | NULL | NULL |
| 59.114.123.73   |   224 | NULL | NULL |
| 12.215.231.131  |   218 | NULL | NULL |
| 68.193.151.157  |   200 | NULL | NULL |
| 98.28.106.213   |   200 | NULL | NULL |
| 122.118.202.29  |   198 | NULL | NULL |
| 67.184.18.83|   196 | NULL | NULL |

There have been fewer than 5 attacks from each of 4515 different IPs.

So for those of you trying to stop this sort of thing by blocking IP addresses, 
don't bother.   

Some of those 203.* and 211.* addresses look suspicious, and perhaps are part 
of the botnet control, but who knows...

I have the complete list of 12,000 IP addresses (and counting at the rate of 
500+ new IP addresses each hour) of this botnet available if that's of any use 
to anyone.

Regards


--- On Fri, 8/8/08, Brad Wood <[EMAIL PROTECTED]> wrote:

> From: Brad Wood <[EMAIL PROTECTED]>
> Subject: Re: SQL injection attack on House of Fusion
> To: "CF-Talk" 
> Date: Friday, August 8, 2008, 4:25 PM
> Yeah, I'm well aware of the near impossibility of ever
> tracking IP address 
> to anything useful, but I'm a person who likes data,
> for within mounds of 
> useless data can be found trends.  Most of all, I'm
> just curious.  Also, I'd 
> like to explore the possibility of reporting compromised
> people to their 
> ISPs.  Some US ISPs do have abuse policies that might cause
> them to pressure 
> their users to clean themselves.  A number of these attacks
> could have come 
> from servers for all we know.  Servers are desirable for
> Trojans due to 
> their fast internet connections and 24/7 uptime.
> 
> ~Brad
> 
> - Original Message - 
> From: "Andy Matthews"
> <[EMAIL PROTECTED]>
> To: "CF-Talk" 
> Sent: Friday, August 08, 2008 3:00 PM
> Subject: RE: SQL injection attack on House of Fusion
> 
> 
> > blocking the IPs would probably stop the attacks, but
> analyzing them is
> > going to be useless. They're either using some
> hacked computer as a proxy,
> > or have some sort of spoofing in place. Unless
> you're really good at
> > forensics, you'll never find their real
> origination point.
> 
> 
> 

~|
Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to 
date
Get the Free Trial
http://ad.doubleclick.net/clk;203748912;27390454;j

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:310572
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

Re: SQL injection attack on House of Fusion

[Terry Ford]

Our site has now seen just over 200,000 attack attempts over the past 48 hours, 
73,000 attack attempts over the past 5 hours.

Not nearly a DOS concern yet, as the acceleration of attacks has started to at 
least flatten a bit over the last 2-3 hours, but we're watching it carefully.

The attacks appear to only effect MSSQL.  MySql seems to be safe from THIS 
attack, but clearly it's time to batten the hatches if you haven't already.

We have 1.2 million pages indexed in Google, where the botnet is extracting 
URLs from.   The level of the attack you are witnessing is likely proportional 
to your presence in google.

The attacks are ONLY against our .CFM pages.   No attacks are being run against 
our obfuscated CFM pages (product.cfm?id=12 => /product/12.html ).

Here's the botnet attack that appears to be hitting us all:

http://www.bloombit.com/Articles/2008/05/ASCII-Encoded-Binary-String-Automated-SQL-Injection.aspx

Decoding the hex of our current attacks and most of them are pointing to 
http://sdo.1000mg.cn/csrss/w.js (malware).   

That is the Asprox botnet, which has beem going through ASP sites for a 
while...  looks like they recruited a bunch of drones, and those drones have 
moved from ASP (verynx attacks) to attack CF.   Pretty ingenious really, 
infecting websites via injection attack in order to infect clients with browser 
vulnerabilities.

The more CF sites that get infected, the more drones that are recruited, and 
the more persistent the attacks become. 

Here's the rewrite I'm using (linux apache) to keep traffic off the app server.

RewriteCond %{QUERY_STRING} .*DECLARE.*
RewriteRule ^(.*)$  violation.htm [nc,L]

Interesting philosophical thought:   

I can't help but believe that the URL rewriting we do over much of our site 
(product.cfm?id=14  appearing as /product/14.html etc etc) has helped reduce 
the attacks significantly.   It seems to me that such URL rewriting is actually 
a very important security tool as we enter a period where botnets start 
targetting .cfm pages.

I plan on increasing our CFM obfuscation over the coming weeks to help hide CF 
from the search engines and automated attacks.   Seems to me that it's a lot 
safer presenting your entire site as HTML to the outside world.

Regards
Terry 

~|
Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to 
date
Get the Free Trial
http://ad.doubleclick.net/clk;203748912;27390454;j

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:310562
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

subscribe

[Terry Ford]

subscribe

~|
Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to 
date
Get the Free Trial
http://ad.doubleclick.net/clk;203748912;27390454;j

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:310552
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: HELP! SQL Injection Attack!

[Terry Ford]

We run a very large linux website and have seen such SQL attacks for months 
now.   

However, this recent wave has *really* picked up over the last 48 hours.It 
began August 6, and we have logged close to 40,000 attack attempts and counting.

Our code is well protected, and this attack appears to target MSSQL, so we are 
not too concerned about the injection.   However, if this attack continues to 
grow, it could quickly become a DOS situation.   Blocking IP addresses is 
futile.   I count hundreds and hundreds of them originating the attack.



~|
Adobe® ColdFusion® 8 software 8 is the most important and dramatic release to 
date
Get the Free Trial
http://ad.doubleclick.net/clk;203748912;27390454;j

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:310454
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

MySQL 4 driver and characterEncoding=ISO-8859

[Terry Ford]

Our site allows international / non-ASCII characters (accents, curly quotes, 
etc) in database data.  We are using MySQL 4.1.

In CF 6.1, it all works great with the following:



  AND, using the MySQL 3 
driver, useUnicode=true&characterEncoding=ISO-8859-1 in the connection string.  
It works.

Now that we've upgraded to CF 8 / the MySQL 4 driver, a lot of our non-ASCII 
characters are getting displayed as boxes or other weird characters.

I suspect there is a very simply fix that I'm overlooking here.  It appears the 
"useUnicode=true&characterEncoding=ISO-8859-1" string used for the MySQL 3 
driver is being ignored by the MySQL 4 driver.

Does anyone have any suggestion on how to get this international data to 
display properly in CF 8 using the MySQL 4 driver? 

Thanks for your thoughts.


~|
Check out the new features and enhancements in the
latest product release - download the "What's New PDF" now
http://download.macromedia.com/pub/labs/coldfusion/cf8_beta_whatsnew_052907.pdf

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:294411
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

CF8 Upgrade Question

[Terry Ford]

Hey... quick question:

We are running CF6.1 and want to upgrade to CF8 (linux).

If we purchase the CF8 upgrade version, does CF6.1 have to be present on the 
machine?  i.e. if we want to install fresh on a new machine in the future, do 
we have to install 6.1 first and then 8, or can we install just the 8 upgrade?  
  I forget how this has worked in the past, but I seem to recall that the old 
version had to be present on the machine at time of "upgrade" install.

Also, do we get a new serial number when we upgrade, or do we use our old CF6 
SN?

Thanks 

~|
Get involved in the latest ColdFusion discussions, product
development sharing, and articles on the Adobe Labs wiki.
http://labs/adobe.com/wiki/index.php/ColdFusion_8

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:294392
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: Facebook integration & auto CFMX form validation

[Terry Ford]

No, this has nothing to do with CFFORM.

CFMX is performing validation on ANY form variables whose names end in "_date".

The data is being POSTed from a 3rd party application.



<< You mean if you use cfform? >>

~|
CF 8 – Scorpio beta now available, 
easily build great internet experiences – Try it now on Labs
http://www.adobe.com/cfusion/entitlement/index.cfm?e=labs_adobecf8_beta

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:280588
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

Facebook integration & auto CFMX form validation

[Terry Ford]

Hey...

It seems that Facebook's new application API requires FORM callbacks that use 
FORM.variable names ending in either _date or _time.

This is triggering CFMX (6.1 at least) to automatically try to validate the 
calls.  This is a bad thing, as the CF developer has no control over the name 
of those variables or the data type.  There is also no ability to intercept the 
FORM and change it, because validation occurs before anything else.

The end result is an HTTP 500 error with:
"Form entries incomplete or invalid.* 1181490277.3472"

If I strip the page down to the string "hello", the above error still occurs.

Has anyone else run into this problem, or does anyone know how to entirely 
disable this automatic FORM validation based on variable name?  I've tried 
playing around with CFERROR, but can't find a way to stop it from aborting the 
application at validation failure.

Thanks

~|
CF 8 – Scorpio beta now available, 
easily build great internet experiences – Try it now on Labs
http://www.adobe.com/cfusion/entitlement/index.cfm?e=labs_adobecf8_beta

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:280585
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: why might sessions end more quickly after changing to linux/B D?

[Terry Ford]

Try turning "J2EE session variables" off, if you're currently using them, and 
restart the server.

We've had problems with J2EE session variables on CFMX6.1 linux in the past and 
currently run with them turned off.

On our server, we get weird stuff like cflock scope=session throwing null 
exceptions and stuff like that when we have J2EE session vars on.  Perhaps what 
you're observing is a symptom of the same problem.



> - The timeout value that's set on the server directly 
> 
> Was 20min, I upped it to 40, no change.
> 
> - The timeout value that you've set in your Application.cfm (if any) 
> 
> It's longer. I do have a search for useragent string for spiders and 
> crawlers and knock them down to 1 second timeouts to preserve memory, 
> but code hasn't changed from previous server and was never an issue 
> before. I'm going to revise that code and see if it does anything.
> 
> - This could also be related to network problems. Are you on a network 
> where IP changes often? 
> 
> No, and it is not only me. Some customers say suddenly their cart will 
> be empty sometimes. Clearly they lost their session. 
> 
> - Are you using cookieless session tracking? (that can get tricky...)
> 
> Cookies.
> 
> I'm clutching at straws here, but was wondering if something to do 
> with vps perhaps and losing the domain/cookie? (And dns propagated 2 
> weeks ago.)
> 
> Thanks for any ideas.

~|
Deploy Web Applications Quickly across the enterprise with ColdFusion MX7 & 
Flex 2
Free Trial 
http://www.adobe.com/products/coldfusion/flex2/?sdid=RVJU

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:280389
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: Reducing phantom sessions?

[Terry Ford]

Ahhh, ok.  That approach makes sense too.  I was trying to shut down session 
mgmt completely for bot requests and had no luck.

Setting a timeout of 1 second doesn't cause any mutex or locking problems does 
it?  i.e. if a bot request creates a 1 second session and runs 10 seconds, does 
the session handler lock up waiting on that request to end?   I would think 
that a session couldn't be deleted while a request of that session is running.  
Perhaps a timeout of 10 or 20 seconds would be safer?

This would be mute, of course, if the 1s timer starts when that initial request 
ends.


>Terry,
>
>The only thing you're doing is setting the timeout for a bot to one second. 
>They still create a session, but the session is destroyed immediately. The 
>whole problem is that sessions are created for each bot hit and exist until 
>the session times out. If you have one bot hitting ten pages, that's ten 
>sessions. If you have ten bots hitting ten pages, that's one hundred 
>sessions. If the session lasts for five minutes, then you can see how it 
>will simply ramp up RAM as the bots come in.
>
>House of Fusion gets tens of thousands of bot hits a day, and by using this 
>technique, we show no loss of memory or performance.
>
>Judith Dinowitz
>Editor-in-Chief: Fusion Authority
>http://www.fusionauthority.com
>
>- Original Message - 
>From: "Terry Ford" <[EMAIL PROTECTED]>
>To: "CF-Talk" 
>Sent: Wednesday, June 06, 2007 10:53 PM
>Subject: Re: Reducing phantom sessions?
>
>
>>

~|
Macromedia ColdFusion MX7
Upgrade to MX7 & experience time-saving features, more productivity.
http://www.adobe.com/products/coldfusion?sdid=RVJW

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:280323
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: Reducing phantom sessions?

[Terry Ford]

It seems that Michael and Ben used a similar solution.

The dynamic setting of sessionmanagement=YES|NO depending on user agent looks 
good in theory.

However, when I switch to that, I get this error all over the place when those 
bot user agents start hitting templates:

"The requested scope session has not been enabled.Before session variables can 
be used, the session state management system must be enabled using the 
CFAPPLICATION tag."

Now I *do* have every reference to a session variable wrapped in an existence 
test.

However, it seems that if you have sessionmanagement=NO then even TESTING for 
the existence of a session variable throws the above error.  i.e. 
isdefined("session.username") throws the above error.  That doesn't make a 
whole lot of sense to me.  

Any ideas how to bypass the error above without having to kludge all sorts of 
stuff around every single test of session variable existence?

Terry


>Terry,
>
>Michael had a similar problem; we get indexed all the time by Google and 
>other bots. He posted the problem and his solutions at
>http://www.blogoffusion.com/index.cfm/2005/11/28/pseudomemory-leak
>
>Judith Dinowitz
>Editor-in-Chief: Fusion Authority
>http://www.fusionauthority.com
>
>- Original Message - 
>
>>

~|
Deploy Web Applications Quickly across the enterprise with ColdFusion MX7 & 
Flex 2
Free Trial 
http://www.adobe.com/products/coldfusion/flex2/?sdid=RVJU

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:280318
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

Reducing phantom sessions?

[Terry Ford]

Phantom Sessions:  These are sessions generally created by bots from Google, 
msn, yahoo, etc.  For every page that they access, a session is created.  Since 
these clients have no cookies, CF creates a new session for every page they 
touch.  Even more annoying are sessions created by CFHTTP calls originating 
from one of your own templates. 

On our website (large, 500,000+ pages in Google), we get Google and buddies 
crawling over our site aggressively all day long.  Add in a handful of visitors 
who disable cookies, and you end up with a huge number of session objects that 
need not exist.

I estimate that about half of all sessions created on our server are useless.  
In peak, that can amount upwards to about 4000+ sessions in a 20 minute span.  
Phantom session objects add overhead, need to be checked for timeouts by the 
system, take up memory, add unnecessary burden to GC, etc.  Is the effect huge? 
 Probably not, because empty session objects are small.  It would be nice to 
limit them though.

Does anyone know of a way to reduce the creation of such "phantom" sessions?  
I'd like CF to be able to say "hello Google, no session object required for 
you".  

Regards
Terry


~|
ColdFusion MX7 by Adobe®
Dyncamically transform webcontent into Adobe PDF with new ColdFusion MX7. 
Free Trial. http://www.adobe.com/products/coldfusion?sdid=RVJV

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:280310
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: CFMX memory usage under load : is this normal?

[Terry Ford]

Thanks. 

I guess the developer in me isn't quite comfortable with so much temporary 
object creation and destruction.  CFMX pools db connections and threads; why 
not pool query objects, instead of instantiating new ones on each request? 

Looks like the JRE creates and destroys new query objects for every single 
request that touches queries, creating lots of memory garbage.  Easy to show 
this by performing queries and printing free allocated mem after each. 

I would think that a more efficient method would be to have a query object pool 
that gets allocated once and used by threads on demand (you'd have to bound 
query result size).  That way you could reuse the same query objects and take 
pressure off the GC.

Surely your application doesn't really need to temporarily allocate and free 
600MB of memory every 30 seconds?  Or is this object-creation-happy use of 
memory cleaned up by the GC actually more efficient than reuse would?

The current usage of memory just seems wrong to me.


>Yes this is normal. Our busy SPARC server has a 2.5 GB JVM and frees
>over 600MB when it does a GC. It does this at roughly 30 second
>intervals or so, like your figures show.
>
>On 6/6/07, Terry Ford <[EMAIL PROTECTED]> wrote:
>
>
>-- 
>mxAjax / CFAjax docs and other useful articles:
>http://www.bifrost.com.au/blog/

~|
Create robust enterprise, web RIAs.
Upgrade & integrate Adobe Coldfusion MX7 with Flex 2
http://www.adobe.com/products/coldfusion/flex2/?sdid=RVJP

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:280200
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

CFMX memory usage under load : is this normal?

[Terry Ford]

I've never actually paid much attention to CFMX's memory usage before, but I 
just started monitoring our memory, and I'm a little unsure whether what we're 
seeing is normal or not.

Setup: CFMX6.1, linux ES4, 800MB Xmx.  Server averages 30 templates/second.
Cached queries set to 0 (same thing happens for different values of cached 
queries anyway).  I'm noticing the same behavior on CF8 pub beta.

Monitoring java.lang.Runtime.getRunTime()'s freeMemory()

Here's a sample of our "free memory" output at 5 second intervals:
493mb
356mb
66mb
66mb
59mb
350mb
57mb

Over the course of a minute, the free memory value is flying all over the 
place.  It dives down to 50mb or so, and then I assume the GC cleans up and 
it's back up in the 350+s.  Then it drops rapidly back to 50mb, and flies back 
up again.

The app is not doing anything that I'd consider memory intensive.  Lots of DB 
queries, but none are that large memory wise.  100 rows here, 2 rows there, but 
done many times each minute.

There is no notable strain on the server (pages load fast, server load avg is 
low), but I'm wondering if this is symptomatic of a memory problem/leak 
somewhere, or if it's completely normal for a loaded server to be allocating 
and freeing memory at such a frantic pace.

Thanks

~|
Deploy Web Applications Quickly across the enterprise with ColdFusion MX7 & 
Flex 2
Free Trial 
http://www.adobe.com/products/coldfusion/flex2/?sdid=RVJU

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:280162
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

Re: Restarting CF Server Manually on Linux

[Terry Ford]

Depending on where you installed CFMX, the path may be different, but it'll be 
something like this:

/opt/coldfusionmx/bin/coldfusion stop
/opt/coldfusionmx/bin/coldfusion start
/etc/rc.d/init.d/httpd restart

You can then do /opt/coldfusionmx/bin/cfstat 1 to watch it process requests.


>I've entered the world of running CF on Linux and it's all new to me. I 
>have a VPS server running CFMX7 and Apache. There was apparently no JDBC 
>connection driver installed for mySQL so I uploaded one and now need to 
>restart CF manually. Can anyone kindly tell me the steps to do that?
>
>Muchos gracias.
>
>-- 
>Best regards,
>
>Chris Montgomery

~|
Upgrade to Adobe ColdFusion MX7
The most significant release in over 10 years. Upgrade & see new features.
http://www.adobe.com/products/coldfusion?sdid=RVJR

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:279897
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Two CFCHART Mysteries...

[Terry Ford]

I've got a couple of mysteries for you CFCHART sleuths...

1.  I can't for the life of me work out how to get the X-axis to display 
integers.  It insists on displaying decimals, even though the query column is 
an integer.

2.  A black line appears on the very left margin of all of my charts.  I'm not 
sure why it's there.  Does anyone have any clue how to get rid of this line?   
I'm running on CFMX6.1 on RHEL4, using the xorg-X11-deprecated-libs package, 
since RH no longer ships with libXp.so.6 any more.

An example illustrating both of these problems is at:

http://www.funtrivia.com/chartdemo.cfm







Thanks for any ideas you might have.

Regards,
Terry

~|
Introducing the Fusion Authority Quarterly Update. 80 pages of hard-hitting,
up-to-date ColdFusion information by your peers, delivered to your door four 
times a year.
http://www.fusionauthority.com/quarterly

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:259388
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: Using CFCHART to generate batch charts?

[Terry Ford]

Thought I might as well answer my own question, for future searchers on this 
topic.

You can actually access the binary object directly by name:








~|
Introducing the Fusion Authority Quarterly Update. 80 pages of hard-hitting,
up-to-date ColdFusion information by your peers, delivered to your door four 
times a year.
http://www.fusionauthority.com/quarterly

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:259385
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

JVM garbage collection pause -- suggestions?

[Terry Ford]

Hi folks...

We're experiencing about 20 second pauses once every few hours.  The problem is 
not db, network, virt memory, or disk related, and the regularity of the 
durations seems to suggest a garbage collection problem. 

We are running on a dualcore Opteron 270.  4GB of memory, machine is dedicated 
to CFMX 6.1 + HTTPD, linux.

I thought that "UseConcMarkSweepGC" was supposed to be low pause, but perhaps 
our large Xmx (we do a ton of memory caching) is causing a problem.  Any 
suggestions on reducing such a GC pause when you require large Xmx's?

Here are the current jvm.arg settings...

java.args=-server -DJINTEGRA_NATIVE_MODE -DJINTEGRA_PREFETCH_ENUMS 
-Xbootclasspath/a:{application.home}/lib/webchartsJava2D.jar
-Djava.awt.graphicsenv=com.gp.java2d.ExHeadlessGraphicsEnvironment -Xmx1000m 
-Dsun.io.useCanonCaches=false -XX:MaxPermSize=128m -XX:PermSize=64m
-XX:+UseConcMarkSweepGC -XX:NewSize=48m

Thanks,
Terry

~|
Introducing the Fusion Authority Quarterly Update. 80 pages of hard-hitting,
up-to-date ColdFusion information by your peers, delivered to your door four 
times a year.
http://www.fusionauthority.com/quarterly

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:259308
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

Re: Second Server

[Terry Ford]

<< Should I just move over the DB or would I be better off Load balancing 
across the two boxes? Or is the something else you'd recommend? >>

Make sure that you also take into consideration how much you care about 
maintenance / administration time.  IMO it's often worth paying a little more 
for a couple of larger boxes than an army of cheaper ones just so that you can 
save time keeping them in order.  All depends on what your priorities are, and 
what you expect your future traffic growth to be.

If you're getting any sort of load, get that db off the CF box.  It's like 
night and day.  

I just upgraded my CF setup from a single dual Xeon to a pair of Opterons, 1 
acting as a mysql server, the other for httpd/cfmx.  Stick a switch between 
them, and watch in amazement.  My experience is that you'll tend to be disk/IO 
bound if you're running your DB on a busy CFMX server.  Fast disks might help, 
but giving the DB a new home is going to be worth it.

If your db server is light loaded, you can pile other tasks on it like 
sendmail, backups, etc and take even more pressure off the CF server.  
Personally I wouldn't consider load balancing at all until a 3rd server was 
required, unless failover is important to you.  Of course, if availability is 
critical then your requirements will vary, in which case you're going to want 
failover of both cf and the db.

~|
Introducing the Fusion Authority Quarterly Update. 80 pages of hard-hitting,
up-to-date ColdFusion information by your peers, delivered to your door four 
times a year.
http://www.fusionauthority.com/quarterly

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:259199
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

Using CFCHART to generate batch charts?

[Terry Ford]

Does anyone know if there is a way to use CFCHART on CFMX to generate a set of 
chart JPEGs and save them to disk?  The idea would be to generate 1000 charts 
in this manner once a week.

This way, simple IMG tags can then be used to link to the charts, and the data 
gathering / CFCHART engine can be entirely bypassed for that week.

There doesn't appear to be a documented way to do this, but perhaps someone has 
worked out how to...

Regards,
Terry

~|
Introducing the Fusion Authority Quarterly Update. 80 pages of hard-hitting,
up-to-date ColdFusion information by your peers, delivered to your door four 
times a year.
http://www.fusionauthority.com/quarterly

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:259198
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Re: CF Server 6.1 Keeps restarting

[Terry Ford]

Funny... because our server has just started doing this too (CF6.1 on linux).

Having run for over a year and a half without a single incident, it has taken 
to restarting itself once every 12 hours or so.

We have made a few changes recently, but it's very difficult to work out what's 
causing this problem.  No error messages anyway.

Does anyone have any ideas what can cause CF to go through the full restart 
process?  Is this a memory issue?  Are we using too much?  Do we need to 
allocate more to the JVM?  Or something else?

Thanks


>Hello,
>
>Our CFserver on Windows 2000 keeps restarting, several times throughout the
>day. The Event log shows " The ColdFusion MX Application Server service for
>the "default" server is restarting."
>
>The Jrun log shows: "2006-09-26 15:16:53 jrISAPI[7544:1664]  jrRecv
>failed[3628]: 10054 Connection reset by peer"
>
>And there are ZERO errors in the application and exception logs. What could
>cause this and how do I track it down? Any suggestions?
>
>Thanks!
>
>BrookD

~|
Introducing the Fusion Authority Quarterly Update. 80 pages of hard-hitting,
up-to-date ColdFusion information by your peers, delivered to your door four 
times a year.
http://www.fusionauthority.com/quarterly

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:256474
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4

Building a Better CF_Accelerate?

[Terry Ford]

Hey folks... 

Haven't been here in a while, but thought I'd pop in and see if anyone has 
developed a more efficient partial-page caching tag in the last couple of years.

We have used CF_Accelerate (by Brandon Purcell) for 3 years now, and it has 
help us wonderuflly.  For those who don't know what it is, it caches parts of a 
page in a struct-based memory structure. 

It has worked well to cache data for various periods of time, essentially 
cutting out all processing / queries for the code around which it appears.  We 
handle upwards of 25 million CFM pageviews (heavy db use) a month on a single 
dual xeon box, hosting mysql and httpd on the same machine, so caching has been 
essential.

One modification we made to speed it up was to use named locks on the 
primary/secondary keys in order to speed up lock waiting that was occurring.

Anyway, I was wondering if anyone has come up with a more optimized/efficient 
CF_Accelerate-esque tag over the past couple of years.  During peak times when 
the server is being hit hard, I've noticed that the CF_Acclerate code itself 
seems to have execution times that, while still fairly fast, could be improved. 
 I suspect it has something to do with structs growing too large and operations 
like isdefined(struct.key.key2) slowing down.

I recall reading a thread somewhere about how CFMX can be either efficient or 
very inefficient when dealing with large structs depending on how the key names 
are distributed.  Perhaps hashing secondary key names would improve the seek 
times on them?  Except perhaps the hashing overhead itself would erase any 
improvement...
 
Regards
Terry

~|
Introducing the Fusion Authority Quarterly Update. 80 pages of hard-hitting,
up-to-date ColdFusion information by your peers, delivered to your door four 
times a year.
http://www.fusionauthority.com/quarterly

Archive: 
http://www.houseoffusion.com/groups/CF-Talk/message.cfm/messageid:256057
Subscription: http://www.houseoffusion.com/groups/CF-Talk/subscribe.cfm
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

Re: CF/mysql Datetime Best Practices

[Terry Ford]

> uh, cf uses's something like "Excel_Time" or "DB2_Time", it's numeric 
> time unit
> is *days* since 31-dec-1899 (it's epoch).

> uh, cf uses's something like "Excel_Time" or "DB2_Time", it's numeric 
> time unit
> is *days* since 31-dec-1899 (it's epoch).

Right.  I was talking about Unix's epoch (POSIX time) for date/time (as my 
first apps were perl on linux), which is manipulated as such:

// get datetime int for storage
dat = DateDiff("s", "January 1 1970 00:00", Now());

// convert an int to whatever date/time format is required
LSdateformat(dateadd("s", dat, createdatetime(1970, 1, 1, 0, 0, 0)),"mmm dd yy")
LStimeformat(dateadd("s", dat, createdatetime(1970, 1, 1, 0, 0, 0))) 

Works fine, and quite fast if you application cache the createdatetime.  I've 
simply liked to rely as little on proprietry database functions as possible.  
Time to convert these to mysql DateTimes and use mysql's datetime functions 
though I reckon.  http://en.wikipedia.org/wiki/Unix_timestamp

~|
Message: http://www.houseoffusion.com/lists.cfm/link=i:4:237850
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

CF/mysql Datetime Best Practices

[Terry Ford]

Hey all,

I've tended to store dates as a simple epoch integer (seconds from Jan, 1970), 
a throwback to perl/tcl/php development days.  While this method is a little 
antiquated, it makes the dates database and software independent.  CF, perl, 
php, database upgrade, whatever... can and will handle these integers fine 
(well, until 2038).  

However, I suspect that some of the mysql date/time datatypes and functions 
offer some serious functionality and time saving improvements.

So I'm just wondering what's considered the current best practice use of date 
and time storage in mysql using CF.  Timestamp?  Datetime?  Neither?  Is using 
epoch times a no-no nowadays, or do other people still use it?  

Regards,
Terry

~|
Message: http://www.houseoffusion.com/lists.cfm/link=i:4:237815
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Avoiding copying by value

[Terry Ford]

Ok, that makes sense (Dan/Kerry/Joe).  In that case, there's no problem  Thanks 
for the clarification!  The lack of a querycopy() function makes the 
interpretation of the assignment operation somewhat ambiguous at first glance.

Terry

>
>All that shows is session.qncache_qn is just a reference to an internal CF
>hash. The structDelete() function is just remove the "qncache_qn" key from
>the session scope--it's not killing the actual reference.
>
>-Dan

~|
Message: http://www.houseoffusion.com/lists.cfm/link=i:4:228495
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Avoiding copying by value

[Terry Ford]

<< If you run the query directly into session.cache_query rather than using 
the intermediate q, you could do exactly what you asked.  For Google and 
other bots/users who don't keep sessions, your check to see if 
session.cache_query would always show it's nonexistant, so it would 
always be rerun. >>

Gotcha, I see what you mean. 



Sounds like a good approach.

The only potential downside is the overhead required to store and then later 
delete the cached queries for clients like Google/yahoo/msn/etc bot sessions.  
Between these guys and clients with no cookies, thousands of unused sessions an 
hour are created, which means that for sessiontimeout minutes that cache_query 
memory will be wasted.   Could always manually delete the variable based on 
user_agent I suppose.  Have never stored queries directly into session scope, 
but would be interested to hear if anyone has any experience with performance 
of such under load vs using local variables.

>If you run the query directly into session.cache_query rather than using 
>the intermediate q, you could do exactly what you asked.  For Google and 
>other bots/users who don't keep sessions, your check to see if 
>session.cache_query would always show it's nonexistant, so it would 
>always be rerun.
>
>--Ben
>
>Terry Ford wrote:
>>

~|
Message: http://www.houseoffusion.com/lists.cfm/link=i:4:228489
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Avoiding copying by value

[Terry Ford]

<< or you could try an iif() in the query="" and based on
isdefined("session.cache_query") use the appropriate query variable >>

Hey, not a bad idea.  Thanks... this is probably the best approach.

Terry


> You could rewrite the loop twice, once for each variable name (probably
>not best), or you could try an iif() in the query="" and based on
>isdefined("session.cache_query") use the appropriate query variable.
>
>~Brad
>
>-Original Message-
>From: Terry Ford [mailto:[EMAIL PROTECTED] 
>Sent: Thursday, January 05, 2006 8:35 AM
>To: CF-Talk
>Subject: Re: Avoiding copying by value
>
>>You can use duplicate, but why not reference session.cache_query?
>>
>>bob
>
>To repeat:
>
>"I know that I could do , but then
>the Google "q" query wouldn't work.  I really want to be able to
>reference either the cached query or the runtime query by the same
>variable name, "q". "

~|
Message: http://www.houseoffusion.com/lists.cfm/link=i:4:228473
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Avoiding copying by value

[Terry Ford]

Hi Joe,

That's what I used to think, too.  But take the following code:





That cfdump works.  This shows that the whole query HAS been copied to Q.

Regards,
Terry



>Hey Terry,
>
>Queries are references types in CF, so it's not an issue.   Example:
>
>
>
>
>
>
>
>
>
>
>
>-Joe
>
>
>On 1/5/06, Terry Ford <[EMAIL PROTECTED]> wrote:
>>

~|
Message: http://www.houseoffusion.com/lists.cfm/link=i:4:228472
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Avoiding copying by value

[Terry Ford]

>You can use duplicate, but why not reference session.cache_query?
>
>bob

To repeat:

"I know that I could do , but then the 
Google "q" query wouldn't work.  I really want to be able to reference either 
the cached query or the runtime query by the same variable name, "q". "

~|
Message: http://www.houseoffusion.com/lists.cfm/link=i:4:228460
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Avoiding copying by value

[Terry Ford]

Here's the scenario:

A small query (session.cache_query) is cached in each user session at login.  
It is referenced in the following manner in a number of templates:








   select...




...


-

Now to my question:  I assume that the line  
copies the query structure to the q variable (instead of passing a pointer) 
because copying by value is what CF likes to do.  In this case that is needless 
overhead.  

Can anyone suggest a way that I can get around the copying of that query by 
value?

I know that I could do , but then the 
Google "q" query wouldn't work.  I really want to be able to reference either 
the cached query or the runtime query by the same variable name, "q".

Thanks,
Terry

~|
Message: http://www.houseoffusion.com/lists.cfm/link=i:4:228454
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Garbage Collection Lag

[Terry Ford]

Thanks -- I'll give that a try.  Have got approx 950 classes in this app, so 
I'll watch closely for out of memory errors wrto MaxPermSize (classes are in 
Perm memory and not Heap, right?)

After fiddling a little further, -XX:+UseConcMarkSweepGC seems to do more harm 
than good on my setup.  The problem is that the normal garbage collection 
appears to run less frequently but for a longer period of time.  So instead of 
frequent 1-2 second delays with UseConcMarkSweepGC , you get infrequent 3-4 
second delays.

The problem is that even 1-2 second delays are nasty when you've got incoming 
requests of 60+ templates/second.  By the time the GC is complete you've got 
~150 incoming requests that have to sort themselves out.   By turning the 
GCConcMarkSweep off you don't get as many interrupting events.  I like to 
completely avoid queued requests if possible.

One last question... is it correct that all cached queries / session / 
application variables are stored in the heap, and not perm memory?  I use a 
*ton* of application variables (tables and parts of tables in structs of arrays 
and arrays of structs and all sorts of other fun configurations) and a fair 
number of cached queries.  Just wondering if the heavy use of memory in 
application variables affects jvm optimization in any way.

Regards,
Terry

>Permanent Generation isn't involved in the garbage collection process at
>all.  The PermGen space is where your JVM stores the meta-data about the
>classes that ColdFusion is creating constantly.  The parameter you need to
>worry about is your NewSize.   One thing I notice that you probably don't
>want to do is set your MaxPermSize so big.  ColdFusion uses a lot of PermGen
>space, but not that much - especially since you only have 512m set aside for
>your total memory size.  And yes, you're right, your JVM size is Xmx +
>PermSize.
>
>Here are some recommendations:
>
>- You have 4GB of RAM on the machine, you should dedicate much more to the
>size of the JVM (assuming this is your production machine).  Perhaps 2GB
>- Set a minimum PermSize.  -XX:PermSize=80m.  
>- I've never seen ColdFusion use more than 120m in PermGen space even for
>massive applications, so set MaxPermSize somewhere in that range
>- Make your New Generation space larger.  Perhaps 25% of your total JVM
>size.
>- The -XX:+UseConcMarkSweepGC setting seems to make no difference in any of
>the tests I've run either so you can not worry about that
>
>Last but not least, if you're using JVM 1.5 look into using jconsole to
>monitor your memory usage.  If you're still on 1.4.2, you're limited to
>seeing just the total JVM usage, but you can still monitor it to see how
>much it's fluctuating during use.  Bottom line is that every application is
>different, so you need to tinker with the size of your memory spaces until
>you get the performance you need.  In the applications I've looked at,
>simply increasing the New Generation size helps tremendously.
>
>-Original Message-
>From: Terry Ford [mailto:[EMAIL PROTECTED] 
>Sent: Wednesday, January 04, 2006 4:39 PM
>To: CF-Talk
>Subject: Garbage Collection Lag
>
>Hey CFers,
>
>As traffic increases and the server has been up for a while I'm starting to
>notice periods where the JVM seems to be taking vacations when it ought to
>be processing templates.  I suspect that this is garbage collection.
>
>The symptom is a cfstat period for 1 or 2 seconds where no templates are
>processed, yet at the same time no templates are queued.  There are no
>queries running on mysql either, so it does look as if the JVM is at work
>and ignoring CFMX.  
>
>It gets pretty bad at times, with such a pause occuring every 30 seconds or
>so.
>
>Any suggestions on how to reduce this garbage collection lag?
>
>Here's what I'm using right now.  dual xeon machine, 4GB RAM, linux.  A ps
>-e at steady state shows the cfusion process using ~650MB of memory (which
>is an unrelated question:  how can CF be using 650MB if Xmx = 512M?  Unless
>the process memory = Xmx + Maxperm?)
>
>java.args=-server -DJINTEGRA_NATIVE_MODE -DJINTEGRA_PREFETCH_ENUMS
>-Xbootclasspath/a:{application.home}/lib/webchartsJava2D.jar
>-Djava.awt.graphicsenv=com.gp.java2d.Ex
>HeadlessGraphicsEnvironment -Xmx512m -Dsun.io.useCanonCaches=false
>-XX:MaxPermSize=256m
>-Djavax.xml.parsers.SAXParserFactory=com.macromedia.crimson.jaxp.SAXParserFa
>cto
>ryImpl
>-Djavax.xml.parsers.DocumentBuilderFactory=com.macromedia.crimson.jaxp.Docum
>entBuilderFactoryImpl -XX:NewSize=48m
>
>Anyway, any ideas would be much appreciated...  does a larger MaxPermSize
>increase or decrease garbage collection lag?  any other params that might
>help out here?  I tried -XX:+UseC

Garbage Collection Lag

[Terry Ford]

Hey CFers,

As traffic increases and the server has been up for a while I'm starting to 
notice periods where the JVM seems to be taking vacations when it ought to be 
processing templates.  I suspect that this is garbage collection.

The symptom is a cfstat period for 1 or 2 seconds where no templates are 
processed, yet at the same time no templates are queued.  There are no queries 
running on mysql either, so it does look as if the JVM is at work and ignoring 
CFMX.  

It gets pretty bad at times, with such a pause occuring every 30 seconds or so.

Any suggestions on how to reduce this garbage collection lag?

Here's what I'm using right now.  dual xeon machine, 4GB RAM, linux.  A ps -e 
at steady state shows the cfusion process using ~650MB of memory (which is an 
unrelated question:  how can CF be using 650MB if Xmx = 512M?  Unless the 
process memory = Xmx + Maxperm?)

java.args=-server -DJINTEGRA_NATIVE_MODE -DJINTEGRA_PREFETCH_ENUMS 
-Xbootclasspath/a:{application.home}/lib/webchartsJava2D.jar 
-Djava.awt.graphicsenv=com.gp.java2d.Ex
HeadlessGraphicsEnvironment -Xmx512m -Dsun.io.useCanonCaches=false 
-XX:MaxPermSize=256m 
-Djavax.xml.parsers.SAXParserFactory=com.macromedia.crimson.jaxp.SAXParserFacto
ryImpl 
-Djavax.xml.parsers.DocumentBuilderFactory=com.macromedia.crimson.jaxp.DocumentBuilderFactoryImpl
 -XX:NewSize=48m

Anyway, any ideas would be much appreciated...  does a larger MaxPermSize 
increase or decrease garbage collection lag?  any other params that might help 
out here?  I tried -XX:+UseConcMarkSweepGC but it didnt seem to help at all.

Thanks

~|
Message: http://www.houseoffusion.com/lists.cfm/link=i:4:228395
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: CFCHART and caching

[Terry Ford]

Ahh, thanks.  Totally missed this line!

name
 
 Page variable name. String. Generates the the graph as binary data and assigns 
it to the specified variable. Suppresses chart display. You can use the name 
value in the cffile tag to write the chart to a file.

Terry

 
>You can write them to disk and do logic when displaying them.
>
>Ade
>
>-Original Message-----
>From: Terry Ford [mailto:[EMAIL PROTECTED]
>Sent: 13 December 2005 14:49
>To: CF-Talk
>Subject: CFCHART and caching
>
>
>Hey folks, quick question -
>
>Haven't used CFCHART much in the past.  I've got a small number of charts
>that I'd like to cache for a long period of time (hours or days).  During
>that time I do NOT want those charts to be regenerated (because they are
>generated based on data that may have since changed and because the runtime
>of the queries involved are expensive).  I want to be guaranteed that the
>charts are generated precisely once.
>
>In CFIDE, "You can specify the maximum number of charts to store in the
>cache. After the cache is full, if a new chart is generated, then the oldest
>chart in the cache will be discarded. Caching of charts allows multiple
>requests for the same chart to be much faster, as the chart is only
>generated once."
>
>So that doesn't solve the problem, because the long-duration charts will
>still be bumped out of the cache if many other newer charts are generated.
>
>Anyone got any ideas about how to save these charts for however long I want?
>Are there any custom tags that will copy the cached file of a chart over to
>another location with a file name of your choice where it will persist for
>as long as you want as an image?
>
>Thanks,
>Terry

~|
Find out how CFTicket can increase your company's customer support 
efficiency by 100%
http://www.houseoffusion.com/banners/view.cfm?bannerid=49

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226929
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

CFCHART and caching

[Terry Ford]

Hey folks, quick question -

Haven't used CFCHART much in the past.  I've got a small number of charts that 
I'd like to cache for a long period of time (hours or days).  During that time 
I do NOT want those charts to be regenerated (because they are generated based 
on data that may have since changed and because the runtime of the queries 
involved are expensive).  I want to be guaranteed that the charts are generated 
precisely once.

In CFIDE, "You can specify the maximum number of charts to store in the cache. 
After the cache is full, if a new chart is generated, then the oldest chart in 
the cache will be discarded. Caching of charts allows multiple requests for the 
same chart to be much faster, as the chart is only generated once."

So that doesn't solve the problem, because the long-duration charts will still 
be bumped out of the cache if many other newer charts are generated.

Anyone got any ideas about how to save these charts for however long I want?  
Are there any custom tags that will copy the cached file of a chart over to 
another location with a file name of your choice where it will persist for as 
long as you want as an image?

Thanks,
Terry


~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226924
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Where did RECORDCOUNT go?

[Terry Ford]

Yep, running 6,1,0,hf59763_611, linux.

"myquery" was just for purposes of this discussion.  In my app the name of the 
query and the reference to it is "qns".  

As a workaround I've started to cache these queries into arrays of structs to 
reduce the load on these CFQUERYs. 

Terry


>I do remember an odd issue w/ CFMX and queries in CFCs under load -
>but that was fixed in 6.1. You said you are on 6.1, right? Outside of
>that - you got me. I do see that the error mentions QNS, and your code
>mentioned MYQUERY, but something that simple would -always- throw an
>error.
>
>On 12/10/05, Terry Ford <[EMAIL PROTECTED]> wrote:
>> >Any possibility you use a UDF that could rewrite the value of myquery?
>> >If you wrap the check in try/catch and cfdump myquery, what do you
>> >see?
>> >
>> >On 12/10/05, Terry Ford <[EMAIL PROTECTED]> wrote:
>> >>
>>
>> Hey Ray,

~|
Discover CFTicket - The leading ColdFusion Help Desk and Trouble 
Ticket application

http://www.houseoffusion.com/banners/view.cfm?bannerid=48

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226728
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Where did RECORDCOUNT go?

[Terry Ford]

>Any possibility you use a UDF that could rewrite the value of myquery?
>If you wrap the check in try/catch and cfdump myquery, what do you
>see?
>
>On 12/10/05, Terry Ford <[EMAIL PROTECTED]> wrote:
>>

Hey Ray,

myquery is not modified in any way whatsoever by any code in any template.  
I'll try wrapping with try/catch for next time it happens.  In trying to debug 
this problem in the past I have used an isdefined("myquery") after the CFQUERY 
and it has returned false.  So it's as if the  is sometimes returning 
no query object whatsoever.

The actual code is:

. do some stuff ...


...


Template "thequery.cfm" contains the cfquery returning myquery, and only the 
cfquery.  I'm not sure if the fact that the query code is being included has 
anything to do with this problem, but I include is as a possibility.

This problem only occcurs on a couple of queries that:

(a) are executed often (tens of thousands of times a day)
(b) have relatively long execution times (up to 2 seconds) when non-cached, and
(c) are cached via cf query caching

In addition, once this error occurs on a query, that identical query will 
return no query object in subsequent executions of that template for some time 
after the initial error occurred.  That leads me to believe that it's a query 
caching issue. 

In addition, this problem seems to occur most often during peak times of the 
day when traffic is the heaviest and a lot of threads are active, leaving me to 
wonder if this problem is the symptom of some obscure race condition that is 
corrupting the query cache.

Regards,
Terry

Here's an instance of the error that I caught in the exception log:

"Error","jrpp-259","12/08/05","04:08:49","cfap","Element RECORDCOUNT is 
undefined in QNS. The specific sequence of files included or processed is: 
/home/trivia/www/private/play.cfm "
coldfusion.runtime.UndefinedElementException: Element RECORDCOUNT is undefined 
in QNS.
at 
coldfusion.runtime.CfJspPage.resolveCanonicalName(CfJspPage.java:1049)
at coldfusion.runtime.CfJspPage._resolve(CfJspPage.java:1004)
at 
coldfusion.runtime.CfJspPage._resolveAndAutoscalarize(CfJspPage.java:)
at cfplay2ecfm1467329454.runPage(/home/trivia/www/private/play.cfm:134)
at coldfusion.runtime.CfJspPage.invoke(CfJspPage.java:147)
at coldfusion.tagext.lang.IncludeTag.doStartTag(IncludeTag.java:357)
at coldfusion.filter.CfincludeFilter.invoke(CfincludeFilter.java:62)
at 
coldfusion.filter.ApplicationFilter.invoke(ApplicationFilter.java:107)
at 
coldfusion.filter.RequestMonitorFilter.invoke(RequestMonitorFilter.java:48)
at coldfusion.filter.PathFilter.invoke(PathFilter.java:80)
at coldfusion.filter.ExceptionFilter.invoke(ExceptionFilter.java:47)
at 
coldfusion.filter.ClientScopePersistenceFilter.invoke(ClientScopePersistenceFilter.java:28)
at coldfusion.filter.BrowserFilter.invoke(BrowserFilter.java:35)
at coldfusion.filter.GlobalsFilter.invoke(GlobalsFilter.java:43)
at coldfusion.filter.DatasourceFilter.invoke(DatasourceFilter.java:22)
at coldfusion.CfmServlet.service(CfmServlet.java:105)
at jrun.servlet.ServletInvoker.invoke(ServletInvoker.java:91)
at jrun.servlet.JRunInvokerChain.invokeNext(JRunInvokerChain.java:42)
at 
jrun.servlet.JRunRequestDispatcher.invoke(JRunRequestDispatcher.java:249)
at 
jrun.servlet.ServletEngineService.dispatch(ServletEngineService.java:527)
at 
jrun.servlet.jrpp.JRunProxyService.invokeRunnable(JRunProxyService.java:192)
at 
jrunx.scheduler.ThreadPool$DownstreamMetrics.invokeRunnable(ThreadPool.java:318)
at 
jrunx.scheduler.ThreadPool$ThreadThrottle.invokeRunnable(ThreadPool.java:426)
at 
jrunx.scheduler.ThreadPool$UpstreamMetrics.invokeRunnable(ThreadPool.java:264)
at jrunx.scheduler.WorkerThread.run(WorkerThread.java:66)

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226726
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Where did RECORDCOUNT go?

[Terry Ford]

select * from blah where yadda yadda yadda...


No records found.
...


This bit of code is executed thousands of times a day (CFMX 6.1 linux).  Most 
of the time it works just fine.  However, occasionally after the server has 
been running for a number of days under load this template will start 
sporadically producing this error:

"Element RECORDCOUNT is undefined in myquery"

Has anyone seen this sort of behavior before and does anyone have any idea why 
myquery.recordcount may suddenly not exist?

Regards,
Terry

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226724
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: sessions and concurrent expiration

[Terry Ford]

<< The request that runs this code must have also run a cfapplication tag
and therefore kept the session alive, no? >>

Yes.  Imagine thought that a thread that is expiring sessions has just 
determined that the session has expired and then gets suspended as the system 
context switches to another thread.  When the session-killer thread finally 
awakes it will kill that session, regardless of any updates to session state 
that the thread with the cfapplication may have made while the session-killer 
thread was asleep.  Thus a potential race condition.  

Still not sure if this is what's really happening (only the Jrun dudes know 
what monitors/synchronization methods are in use -- perhaps this scenario is 
protected against), but it's the only explanation I can find for the 
disappearance of a session variable right after isdefined has said it's there.  
We often have over 4000 active sessions under load at once too, so the 
opportunity for freak race conditions to show up is more likely. 


> The request that runs this code must have also run a cfapplication 
> tag
> and therefore kept the session alive, no?
> 
> On 12/10/05, Terry Ford <[EMAIL PROTECTED]> wrote:
> > This is what appears to be happening:
> >
> > The thread with this code switches out precisely after the 
> structkeyexists() evaluates true.  The thread that clears inactive 
> sessions becomes active right after it has previously evaluated this 
> session as expiring, so it deletes the session.  The original thread 
> becomes active again, and now has no session, so the second part of 
> the CFIF "session.acceess eq 1" evaluates as an error.
> >
> > If this is the case then I don't think the race condition could even 
> be prevented by CFLOCKing the CFIF block, as the session-clearing 
> thread would ignore it.  It probably hardly ever comes up because the 
> chance of someone suddenly waking up and executing a template that 
> interweaves in that fashion at the exact millisecond that the session 
> times out is rather small.  Cute, though.
> 
> --
> CFAJAX docs and other useful articles:
> http://jr-holmes.coldfusionjournal.
com/

~|
Find out how CFTicket can increase your company's customer support 
efficiency by 100%
http://www.houseoffusion.com/banners/view.cfm?bannerid=49

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226722
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: sessions and concurrent expiration

[Terry Ford]

This is what appears to be happening:

The thread with this code switches out precisely after the structkeyexists() 
evaluates true.  The thread that clears inactive sessions becomes active right 
after it has previously evaluated this session as expiring, so it deletes the 
session.  The original thread becomes active again, and now has no session, so 
the second part of the CFIF "session.acceess eq 1" evaluates as an error. 

If this is the case then I don't think the race condition could even be 
prevented by CFLOCKing the CFIF block, as the session-clearing thread would 
ignore it.  It probably hardly ever comes up because the chance of someone 
suddenly waking up and executing a template that interweaves in that fashion at 
the exact millisecond that the session times out is rather small.  Cute, though.


>Maybe try:
>
>
>
>-- 
>Damien McKenna - Web Developer - [EMAIL PROTECTED]
>The Limu Company - http://www.thelimucompany.com/ - 407-804-1014
>#include 

~|
Discover CFTicket - The leading ColdFusion Help Desk and Trouble 
Ticket application

http://www.houseoffusion.com/banners/view.cfm?bannerid=48

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226712
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

sessions and concurrent expiration

[Terry Ford]

Just curious... does anyone know if it is it possible for a session to 
automatically expire while the user of that session is concurrently processing 
a template?   Does a race condition exist here?

I have a line that is executed frequently:



99.9% of the  time this works as it should.  Now under load once in a very 
blue moon I get this error in my error log: "Element ACCESS is undefined in 
SESSION."  

Given that none of my code deletes sessions or the "access" element, the only 
way this error could be occuring appears to be if the session is actually 
expired between the isdefined("session.access") and session.access eq 1 
statements.

Is that possible?  Can the automatic session expirer chop a session that has 
just become active in a concurrent thread?   Or is a session protected from 
expiration the instant that the session starts executing a template, in which 
case I'm observing some totally unrelated problem?

Regards,
Terry

~|
Find out how CFTicket can increase your company's customer support 
efficiency by 100%
http://www.houseoffusion.com/banners/view.cfm?bannerid=49

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226664
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Cached queries & memory usage

[Terry Ford]

<< I'm a huge fan of the "build well, optimize later" approach to
programming.  Build your app as well as you can, using the simplest
means reasonable, and be confident that the app is more than good
enough until it's objectively demonstrated not to be.   >>

Yep, that's the approach I take too.  I'm constantly looking to improve 
application response once something new is up and running.  One of the best 
ways of determining how something will act in a real environment is to let it 
run in that real environment and see how it does act, and then optimize from 
there.  

I suppose that's why query caching has always bothered me as far as 
optimization is concerned, because it really is a great bit of functionality 
but is largely unobservable in realtime testing.  

So many other aspects of the process are measurable, such as using cfstat, 
observing mysql query times, timing iterative processes, etc.  Query caching is 
one of CF's few enigmas, particularly when it comes to its relatively 
unpredictable and unobservable use of memory.   If you are using your own 
caching mechanisms or storing lots of data in application/session scopes in 
addition to query caching, you might as well throw a dart at a board to 
determine how much memory your CF cached queries are really using.

Terry


>If you really care that much, you're going to be doing some pretty
>in-depth load testing anyway, so you can vary the different cache
>amounts (query cache, other manual caches) and get app-specific
>results.  And even then, it's still irrelevant what the absolute
>efficiency of the query cache is.  It's a question of "good enough"
>that can only be answered by testing your specific app.  No
>preexisting data will be meaningful, because your app will be
>different than whatever app was used to generate the data.
>
>I'm a huge fan of the "build well, optimize later" approach to
>programming.  Build your app as well as you can, using the simplest
>means reasonable, and be confident that the app is more than good
>enough until it's objectively demonstrated not to be.  So use the
>query cache until it's been proven insufficient, and then find a more
>efficient way to do it.  But only worry about it if it's been
>DEMONSTRATED insufficient.
>
>cheers,
>barneyb
>
>On 12/7/05, Terry Ford <[EMAIL PROTECTED]> wrote:
>
>--
>Barney Boisvert
>[EMAIL PROTECTED]
>360.319.6145
>http://www.barneyb.com/
>
>Got Gmail? I have 100 invites.

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226523
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Cached queries & memory usage

[Terry Ford]

<< Honestly, you really ought not to care.  >>

The reason I care is that when you're using a large number of cached queries of 
differing (sometimes variable) sizes in a large, highly loaded, dynamic 
environment, it's not at all easy to know how many queries you ought to cache, 
nor what sort of durations you should place on each one.   It would be quite 
insightful to know what the average size of a cached query over a period of 
time in the cache was, to better tune the "number of queries" option based on 
the amount of memory you're willing to dedicate to that task.

As far as the question of performance, it is very important to understand the 
strengths and weaknesses of CF and Java's data types and caching mechanisms 
precisely so you can better choose between using one of them or rolling your 
own.  Anyone truly interested in high performance and optimization "ought to 
care".  

Regards,
Terry 


> Honestly, you really ought not to care.  I'd say it's safe to assume
> that MM made it operate as quickly and efficiently as possible.  If
> the functionality the built-in cache uses is sufficient for your
> needs, it's very likely it'll be faster than anything you can build 
> in
> CFML (simply because the built-in stuff is written directly in Java). 
> 
> If the functionality isn't sufficient, then you have to write your
> own.
> 
> CF provides no means to check the sizes of in-memory objects (and I
> don't thinkg Java does either - except perhaps via JVM monitoring
> tools), so it's pretty unlikely that you'll be able to do that either
> way.
> 
> Since queries are cached based on exact SQL matching, I'd consider it
> a safe wager that the cache does something like this:
> 
> function runCachedQuery(sql) {
  
> var key = hash(sql);
  
> if (! queryCache.has(key)) {

> queryCache.put(key, runQuery(sql));
  
> }
  
> return queryCache.get(key);
> }
> 
> cheers,
> barneyb
> 
> On 12/7/05, Terry Ford <[EMAIL PROTECTED]> wrote:
> > Has anyone found a way to get access to the underlying java 
> object(s) that CFMX uses for query caching?
> >
> > It would be nice to be able to see precisely (a) how many queries 
> have been cached and (b) how much memory each one is using at any 
> point in time.  Additionally, it would be nice to know precisely how 
> the access times for a query scales as the number of cached queries 
> increase (a hash?), and how those access times compare to self-caching 
> in arrays or structs.
> >
> > Regards,
> > Terry
> >
> > 

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226519
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Cached queries & memory usage

[Terry Ford]

Has anyone found a way to get access to the underlying java object(s) that CFMX 
uses for query caching?

It would be nice to be able to see precisely (a) how many queries have been 
cached and (b) how much memory each one is using at any point in time.  
Additionally, it would be nice to know precisely how the access times for a 
query scales as the number of cached queries increase (a hash?), and how those 
access times compare to self-caching in arrays or structs.

Regards,
Terry

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226507
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: pseudo-memory leak

[Terry Ford]

Well if you're going to go to all the trouble (and performance hit) of building 
your own session scope and copying structs and stuff, you might as well just 
wave the white flag and let the bots use session management ;)

Terry


>More info
>
>You can't do this - at least not this way You can referenced something
>named "session" in the variables scope (variables.session.blah), but when
>you try to use the order of precedence in your favor to evaluate (var="#session#"">) you get an empty structure...
>
>Of course you could "setup" the session and copy it to another scope with
>each request.. using it for storage only... in that case you would end up
>with something like this.
>
>
>
>
>
>   clientmanagement="Yes">
>
>   
>   
>
>
>
>
>   
>
>   
>  (or 
> structcopy or
>duplicate)
>   
>
>   
>   
>   
>   
>
>
>
>
>
>
>
>Then, everywhere you currently use "session" you would use
>"localcopyofsession".
>
>This would cause the bots to all share the same session values
>
>-Mark
>
>
>-Original Message-
>From: Mark A Kruger [mailto:[EMAIL PROTECTED]
>Sent: Wednesday, December 07, 2005 3:22 PM
>To: CF-Talk
>Subject: RE: pseudo-memory leak
>
>
>Regarding my "application name" approach - you could then try the following
>in "botapplication"
>
>
>session = structnew();
>
>session.user = 0
>session.id = 0
>
>. etc
>
>It would actually be "variables.session"  - but it would allow your code to
>work (assuming you can come up with defaults for all your session vars.)

~|
Find out how CFTicket can increase your company's customer support 
efficiency by 100%
http://www.houseoffusion.com/banners/view.cfm?bannerid=49

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226491
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: pseudo-memory leak

[Terry Ford]

Hi Mark,

Yes, that is understood and is precisely what I'm doing.

The issue is that any user of "botApplication" who hits any template containing 
a reference to a session variable (*even if protected within isdefined() 
checks*) will receive an error.  At least that is the behavior with CFMX 6.1.

Regards,
Terry


>Terry,
>
>Session management is set at the "application" level. It's not different for
>each visitor - it is the same for all pages governed by the
>"application.cfm" page.  Try this approach.
>
>
>
>
>   
>
>
>   
>
>
>
>
>this code will in effect create 2 applications - botApplication and
>regularuserapplication - one with mangement "ON" and one with mangement
>"OFF".
>
>-Mark
>
>
>-Original Message-
>From: Terry Ford [mailto:[EMAIL PROTECTED]
>Sent: Wednesday, December 07, 2005 12:57 PM
>To: CF-Talk
>Subject: Re: pseudo-memory leak
>
>
>Correct, that's what I'm trying to do -- disable session management entirely
>for bots so that thousands of sessions dont get needlessly created and
>destroyed every 15 minutes.
>
>Just to put this in perspective as to why I'm experimenting with this,
>Google has over 500,000 pages from my site indexed.  Add in Google Adwords
>on a large number of those pages, and Google is literally hammering the site
>24 hrs a day.  This is great for driving traffic, but the overhead of
>maintaining state for each one of these bot requests seems wasteful to me.
>
>The problem is that as soon as you disable session management, all pages
>that reference session variables (even inside of scope existence checks)
>throw errors.
>
>So it doesn't look like there's a way to me of disabling session management
>for certain cgi.http_user_agents, yet having it enabled for non-bots.

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226488
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: pseudo-memory leak

[Terry Ford]

Correct, that's what I'm trying to do -- disable session management entirely 
for bots so that thousands of sessions dont get needlessly created and 
destroyed every 15 minutes.   

Just to put this in perspective as to why I'm experimenting with this, Google 
has over 500,000 pages from my site indexed.  Add in Google Adwords on a large 
number of those pages, and Google is literally hammering the site 24 hrs a day. 
 This is great for driving traffic, but the overhead of maintaining state for 
each one of these bot requests seems wasteful to me.

The problem is that as soon as you disable session management, all pages that 
reference session variables (even inside of scope existence checks) throw 
errors.

So it doesn't look like there's a way to me of disabling session management for 
certain cgi.http_user_agents, yet having it enabled for non-bots.


>ugh, you have sessionmanagement="no"
>
>DK
>
>On 12/7/05, Terry Ford <[EMAIL PROTECTED]> wrote:
>>

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226473
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: pseudo-memory leak

[Terry Ford]

Nope, still doesn't work.  Even though the "session.pid" test is never reached, 
the error is still thrown.  Try the following code:





An error is thrown by the above statement each time



 

>Add in checker code to see if a session exists before using the session var.
>IsDefined('session')
>
>
>>

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226469
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: pseudo-memory leak

[Terry Ford]

Just tried your approach (sessionmanagement=no for bots) but every bot that 
hits a page that references a session variable (even in a check) then throws an 
error:  

"Before session variables can be used, the session state management system must 
be enabled using the CFAPPLICATION tag."  
  
Any idea on how to get around this error so that the bots can see content?  It 
would be nice to at least allow the bots to index something other than an error 
message :)

Terry


> Interesting topic.
> 
> We gave up using client variables a while back when they started 
> severely impacting performance under load.   It's incredibly important 
> to avoid client variables on systems under load.  When we switched 
> from client vars to a cookie (hashed user ID & password combo) and 
> used sessions exclusively, the site performance improved significantly.
> 
> 
> The problem you outline seems to be a problem for session variables 
> too. Using the runtime.SessionTracker, CFMX reports that there are 
> currently 3900 open sessions (20 min timeout) on our server.  Probably 
> about 1/3 to 1/2 of them are either users who dont have cookies turned 
> on (and therefore create a new session each page), or are bots.  It's 
> unclear how much of a performance hit this needless creation and 
> destruction of sessions causes, but it certainly isn't optimal. 
> 
> Thanks for the cfapplication switching idea.  I'll give it a shot and 
> see if the number of active sessions drops.
> 
> I'd be interested in hearing if anyone has any ideas about stopping 
> sessions from cookieless people from being created (that cant be 
> identified as being bots) in the first place.  I've got a feeling 
> there is no solution for this one though.
> 
> Regards,
> Terry
> 
> 
> 
> > Sorry, meant to post the whole thing. :)
> > 
 
> 
> > For the last few weeks I've been having some problems with House of 
> 
> > Fusion. The memory for the JRun.exe has been going through the roof 
> 
> > and I didn't know why. The code was tight, nothing had really 
> changed 
> > on the site, so what was up? The answer was Yahoo. 
> > In the last 3 weeks Yahoo has ramped up their indexing of sites. For 
> a 
> > site as large as House of Fusion, this can take quite a bit of time. 
> 
> > I've logged 2-4 yahoo bot hits per second at some times. 
> > So how was yahoo the problem? Because of client variables. Not DB 
> > client variables and not even the dreaded registry client variables. 
> 
> > Just simple cookie based client variables. It seems that when a 
> client 
> > variable is set, a memory structure is also set for CF. Now each bot 
> 
> > hit is assumed to be it's own session as it does not accept cookies. 
> 
> > This mean each bot hit generates a memory structure of about 1k. Now 
> 
> > this is not really a lot, but when you have a few 10's of thousands 
> of 
> > hits from bots a day, it adds up. 
> > I'm still waiting on word from Macromedia as to when a client memory 
> 
> > structure times out, but this seems to be the issue. 
> > So what's the solution? There are 4.
> > 1. Increase your ram. If you can do this, then ramp up your memory 
> as 
> > high as you can. This is not a perfect solution but it saves 
> throwing 
> > time at the problem and gives you a 'buffer' against problems of 
> this 
> > sort.
> > 2. Set a robots.txt with a Crawl-delay setting. Mine is set to 1 
> > second but you can set yours to something higher
> > 3. set a different cfapplication for the most common bots. I use a 
> > simple regular expression to find key words that only exist in 
> bots:
> >  REFindNoCase('Slurp|Goo> 
> glebot|BecomeBot|msnbot|Mediapartners-Google|ZyBorg|RufusBot|EMonitor', 
> 
> > cgi.http_user_agent)>
> >  > sessionmanagement="no" setclientcookies="no" setdomaincookies="no" 
> > clientstorage="Cookie">
> > 
> >  > sessionmanagement="no" setclientcookies="yes" setdomaincookies="no" 
> 
> > clientstorage="Cookie">
> > 
> > This will make sure that a client structure is NOT created for one 
> of 
> > these bots.
> > 4. Use the same regex to clean out the client structure after the 
> bot 
> > finishes the page. Use structclear(client) to remove the data in the 
> 
> > onRequestEnd.cfm, the onRequestEnd method of the application.cfc or 
> in 
> > the template itself.
> > Bottom line is that while bots are great for indexing your content, 
> 
> > they can cause havoc on your system when a lot of memory is assigned 
> 
> > to what is essentially a 'dead session'. 
> > 
> > http://www.blogoffusion.com/index.
cfm/2005/11/28/pseudomemory-leak

~|
Discover CFTicket - The leading ColdFusion Help Desk and Trouble 
Ticket application

http://www.houseoffusion.com/banners/view.cfm?bannerid=48

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226462
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http

Re: pseudo-memory leak

[Terry Ford]

Interesting topic.

We gave up using client variables a while back when they started severely 
impacting performance under load.   It's incredibly important to avoid client 
variables on systems under load.  When we switched from client vars to a cookie 
(hashed user ID & password combo) and used sessions exclusively, the site 
performance improved significantly.

The problem you outline seems to be a problem for session variables too. Using 
the runtime.SessionTracker, CFMX reports that there are currently 3900 open 
sessions (20 min timeout) on our server.  Probably about 1/3 to 1/2 of them are 
either users who dont have cookies turned on (and therefore create a new 
session each page), or are bots.  It's unclear how much of a performance hit 
this needless creation and destruction of sessions causes, but it certainly 
isn't optimal. 

Thanks for the cfapplication switching idea.  I'll give it a shot and see if 
the number of active sessions drops.

I'd be interested in hearing if anyone has any ideas about stopping sessions 
from cookieless people from being created (that cant be identified as being 
bots) in the first place.  I've got a feeling there is no solution for this one 
though.

Regards,
Terry



> Sorry, meant to post the whole thing. :)
> 
 
> For the last few weeks I've been having some problems with House of 
> Fusion. The memory for the JRun.exe has been going through the roof 
> and I didn't know why. The code was tight, nothing had really changed 
> on the site, so what was up? The answer was Yahoo. 
> In the last 3 weeks Yahoo has ramped up their indexing of sites. For a 
> site as large as House of Fusion, this can take quite a bit of time. 
> I've logged 2-4 yahoo bot hits per second at some times. 
> So how was yahoo the problem? Because of client variables. Not DB 
> client variables and not even the dreaded registry client variables. 
> Just simple cookie based client variables. It seems that when a client 
> variable is set, a memory structure is also set for CF. Now each bot 
> hit is assumed to be it's own session as it does not accept cookies. 
> This mean each bot hit generates a memory structure of about 1k. Now 
> this is not really a lot, but when you have a few 10's of thousands of 
> hits from bots a day, it adds up. 
> I'm still waiting on word from Macromedia as to when a client memory 
> structure times out, but this seems to be the issue. 
> So what's the solution? There are 4.
> 1. Increase your ram. If you can do this, then ramp up your memory as 
> high as you can. This is not a perfect solution but it saves throwing 
> time at the problem and gives you a 'buffer' against problems of this 
> sort.
> 2. Set a robots.txt with a Crawl-delay setting. Mine is set to 1 
> second but you can set yours to something higher
> 3. set a different cfapplication for the most common bots. I use a 
> simple regular expression to find key words that only exist in bots:
>  
glebot|BecomeBot|msnbot|Mediapartners-Google|ZyBorg|RufusBot|EMonitor', 
> cgi.http_user_agent)>
>  sessionmanagement="no" setclientcookies="no" setdomaincookies="no" 
> clientstorage="Cookie">
> 
>  sessionmanagement="no" setclientcookies="yes" setdomaincookies="no" 
> clientstorage="Cookie">
> 
> This will make sure that a client structure is NOT created for one of 
> these bots.
> 4. Use the same regex to clean out the client structure after the bot 
> finishes the page. Use structclear(client) to remove the data in the 
> onRequestEnd.cfm, the onRequestEnd method of the application.cfc or in 
> the template itself.
> Bottom line is that while bots are great for indexing your content, 
> they can cause havoc on your system when a lot of memory is assigned 
> to what is essentially a 'dead session'. 
> 
> http://www.blogoffusion.com/index.cfm/2005/11/28/pseudomemory-leak

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226457
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Linux CFMX 6.1 / JVM crash revisited

[Terry Ford]

I did try installing the newest sun JDK (1.5_06) but get these errors on 
startup:

Does anyone have any tips on getting the 1.5 JDK to work with CFMX?

Thanks

--

12/06 18:54:41 user CFCServlet: destroy
12/06 18:54:41 user GraphServlet: destroy
12/06 18:54:41 user FlashGateway: destroy
Exception thrown by error-handling template:
coldfusion.server.ServiceFactory$ServiceNotAvailableException: The Runtime 
service is not available.
at 
coldfusion.server.ServiceFactory.getRuntimeService(ServiceFactory.java:107)
at 
coldfusion.filter.ExceptionFilter.handleException(ExceptionFilter.java:98)
at coldfusion.filter.ExceptionFilter.invoke(ExceptionFilter.java:61)
at 
coldfusion.filter.ClientScopePersistenceFilter.invoke(ClientScopePersistenceFilter.java:28)
at coldfusion.filter.BrowserFilter.invoke(BrowserFilter.java:35)
at coldfusion.filter.GlobalsFilter.invoke(GlobalsFilter.java:43)
at coldfusion.filter.DatasourceFilter.invoke(DatasourceFilter.java:22)
at coldfusion.CfmServlet.service(CfmServlet.java:105)
at jrun.servlet.ServletInvoker.invoke(ServletInvoker.java:91)
at jrun.servlet.JRunInvokerChain.invokeNext(JRunInvokerChain.java:42)
at 
jrun.servlet.JRunRequestDispatcher.invoke(JRunRequestDispatcher.java:249)
at 
jrun.servlet.ServletEngineService.dispatch(ServletEngineService.java:527)
at 
jrun.servlet.jrpp.JRunProxyService.invokeRunnable(JRunProxyService.java:192)
at 
jrunx.scheduler.ThreadPool$DownstreamMetrics.invokeRunnable(ThreadPool.java:318)
at 
jrunx.scheduler.ThreadPool$ThreadThrottle.invokeRunnable(ThreadPool.java:426)
at 
jrunx.scheduler.ThreadPool$UpstreamMetrics.invokeRunnable(ThreadPool.java:264)
at jrunx.scheduler.WorkerThread.run(WorkerThread.java:66)
12/06 18:54:41 error The ClientScope service is not available.
coldfusion.server.ServiceFactory$ServiceNotAvailableException: The ClientScope 
service is not available.
at 
coldfusion.server.ServiceFactory.getClientScopeService(ServiceFactory.java:125)
at 
coldfusion.filter.ClientScopePersistenceFilter.invoke(ClientScopePersistenceFilter.java:31)
at coldfusion.filter.BrowserFilter.invoke(BrowserFilter.java:35)
at coldfusion.filter.GlobalsFilter.invoke(GlobalsFilter.java:43)
at coldfusion.filter.DatasourceFilter.invoke(DatasourceFilter.java:22)
at coldfusion.CfmServlet.service(CfmServlet.java:105)
at jrun.servlet.ServletInvoker.invoke(ServletInvoker.java:91)
at jrun.servlet.JRunInvokerChain.invokeNext(JRunInvokerChain.java:42)
at 
jrun.servlet.JRunRequestDispatcher.invoke(JRunRequestDispatcher.java:249)
at 
jrun.servlet.ServletEngineService.dispatch(ServletEngineService.java:527)
at 
jrun.servlet.jrpp.JRunProxyService.invokeRunnable(JRunProxyService.java:192)
at 
jrunx.scheduler.ThreadPool$DownstreamMetrics.invokeRunnable(ThreadPool.java:318)
at 
jrunx.scheduler.ThreadPool$ThreadThrottle.invokeRunnable(ThreadPool.java:426)
at 
jrunx.scheduler.ThreadPool$UpstreamMetrics.invokeRunnable(ThreadPool.java:264)
at jrunx.scheduler.WorkerThread.run(WorkerThread.java:66)

12/06 18:54:41 error Exception occurred invoking custom error page 
/WEB-INF/exception/java/lang/Exception.cfm
coldfusion.server.ServiceFactory$ServiceNotAvailableException: The Debugging 
service is not available.
at 
coldfusion.server.ServiceFactory.getDebuggingService(ServiceFactory.java:98)
at coldfusion.CfmServlet.getFilterChain(CfmServlet.java:72)
at coldfusion.CfmServlet.service(CfmServlet.java:101)
at jrun.servlet.ServletInvoker.invoke(ServletInvoker.java:91)
at jrun.servlet.JRunInvokerChain.invokeNext(JRunInvokerChain.java:42)
at 
jrun.servlet.JRunRequestDispatcher.invokeNext(JRunRequestDispatcher.java:447)
at 
jrun.servlet.JRunRequestDispatcher.include(JRunRequestDispatcher.java:379)
at jrun.servlet.JRunResponse.sendErrorPage(JRunResponse.java:672)
at jrun.servlet.JRunResponse.sendError(JRunResponse.java:515)
at 
jrun.servlet.JRunRequestDispatcher.invoke(JRunRequestDispatcher.java:291)
at 
jrun.servlet.ServletEngineService.dispatch(ServletEngineService.java:527)
at 
jrun.servlet.jrpp.JRunProxyService.invokeRunnable(JRunProxyService.java:192)
at 
jrunx.scheduler.ThreadPool$DownstreamMetrics.invokeRunnable(ThreadPool.java:318)
at 
jrunx.scheduler.ThreadPool$ThreadThrottle.invokeRunnable(ThreadPool.java:426)
at 
jrunx.scheduler.ThreadPool$UpstreamMetrics.invokeRunnable(ThreadPool.java:264)
at jrunx.scheduler.WorkerThread.run(WorkerThread.java:66)

Starting Macromedia JRun 4 (Build 75991), default server
[1]java.lang.NoSuchMethodException: setURL
at 
com.sun.jmx.mbeanserver.StandardMetaDataImpl.forbidInvokeGetterSetter(StandardMetaDataImpl.java:493)
at 
com.sun.jmx.mbeanserver.StandardMetaDat

Linux CFMX 6.1 / JVM crash revisited

[Terry Ford]

Hey folks,

Revisiting a thread from a couple of years ago that I never did get an answer 
to back then, but which was able to be worked around.  Just wondering if anyone 
now has an answer :)

The problem:  On linux, CFMX 6.1, sun JVM 1.4.2, cfserver under heavy load 
crashes randomly every couple of days due to a "hotspot error" reported by the 
JVM.  We determined that it was a bug in the JVM's hotspot compiler, and 
therefore determined that using the "-Xint" JVM flag (disabling it) could stop 
the crashes.

Good news: the server has run fine since in this "-Xint" mode.   Bad news: the 
performance hit of this workaround is less than optimal.  

I'm just wondering if anyone ever found a true fix for this problem (some 
combination of JVM flags?  some updated sun JVM?).  

Alternatively, is there a better JVM (that works!) for CFMX on linux now?  I've 
tried swapping out different JVMs (BEA's, IBM's) through the years with little 
success, as CFMX complains about not being able to find some function or other.

Any help with finding a stable, high performance JVM for CFMX on linux would be 
much appreciated. 

Regards,
Terry

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:226349
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: client variables and subdomains

[Terry Ford]

Hey Matt,

Hmm, setdomaincookies=YES doesn't seem to help either.  Server still creating 
different instances for www.domain.com and .domain.com.  

My use of client variables is actually pretty minor -- I just use a couple to 
store an integer key to a user table, and an encrypted password.  When a user 
returns to the site, I use those two client.vars to setup session info, tailor 
output, etc.  I suppose I could just use CFCOOKIEs for this and have greater 
control over how the cookies are set.  

Re: session variables and CFMX.  As I understand them, session variables in 
CFMX persist only as long as the browser is open, so wouldn't they be ill 
suited for the use outlined above, that of identifying returning visitors who 
may have been away for months?

I'll check out the link you provided.  

Thanks,
Terry

>it is the default behavior.  You are supposed to be able to fix this
>by adding "setdomaincookies="yes" in /Application.cfm.
>
>This is one of the server side code gotchas I was alluding to in your
>other thread.   I have to rely on client vars for practically
>*everything* insofar as state info and I just friggin' hate the grief
>that goes with client vars in cookies.  If you've got MX you can use
>session vars and save yourself a TON of grief.
>
>Have you checked out the CFMX Coding Standards?  This page seems pertinent:
>
>http://livedocs.macromedia.com/wtg/public/coding_standards/structure.html
>
>-- 
>--mattRobertson--
>Janitor, MSB Web Systems
>mysecretbase.com

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:205999
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

client variables and subdomains

[Terry Ford]

To clarify --

I'm simply trying to set a client variable (storage type "cookie"), so that it 
gets set and can be read from any template under *.domain.com.  So I'd like 
www.domain.com, hello.domain.com, and domain.com all to be able to read/set the 
same client variables for a user.  If the user goes to domain.com one day and 
www.domain.com the next, their experience ought to be the same.  They should 
not have to log in again.

However, at the moment it seems that separate instances of the client variables 
are created under "www.domain.com" and "domain.com".  If you login to the site 
from a template under "http://domain.com";, the second you switch to 
"http://www.domain.com"; you must login again.  Is this actually supposed to be 
the default behavior of client variables when stored as cookies, or is 
something wrong? 

Thanks,
Terry

> Thanks for the responses.  Lots of good food for thought.
> 
> The reason that I initially went with database-bound client vars was 
> due to an early CFMX bug whereby cookies weren't getting set correctly 
> in all browsers.  I believe it was fixed in CFMX 6.1, but since 
> everything was smoothly running on the setup at the time didn't bother 
> changing it then.   
> 
> I made the change back to client cookies today, and it appears to be 
> working largely as intended.  
> 
> However, cookies still seem to behave erratically.  For example, I 
> find a client var set (setdomaincookies=yes) from a page at "domain.
> com/login.cfm" is usually also found on a page at "www.domain.
> com/login.cfm" in IE,  but not in Firefox.  Firefox isn't 
> acknowledging "www.domain.com" and "domain.com" as equivalent for 
> purposes of client cookies.
> 
> To make things even more confusing, if I delete a client variable in 
> www.domain.com/test.cfm then re-create it in www.domain.com/test2.cfm, 
> sometimes IE doesn't even acknowledge that it has been set.  IE seems 
> to get confused and will refuse to set the cookie correctly until it's 
> either rebooted or you use "clear all cookies" in the browser.   
> 
> I honestly am not sure what the deal is, but I'm finding it very 
> frustrating trying to get multiple browsers to react consistently to 
> "cookie" client variables.  Right now it seems to me that client 
> variables (type=cookie), when set, do one of three things in the same 
> browser, depending on how the browser is feeling at the time:
> 
> - cli var gets set correctly, and is acknowledged under *.domain.com
> - cli var gets set correctly, but is only acknowledged under current.
> domain.com
> - doesn't get set at all. browser entirely ignores it.  It tends to 
> exhibit this behavior if the client variable was just recently deleted 
> and then re-established (such as a login after a logout).
> 
> These behaviors disappear the INSTANT I turn client storage from 
> "cookie" to "database". 
> 
> If anyone has any insight into what the problem may be, please let me 
> know.  As I'm on update 6.1, I didn't think I needed the 
> "setdomaincookies" hot fix, which is listed on the MM site as an MX 6.
> 0 fix.  Whose fault is this erratic behavior... CF's, IE's, or mine?
> 
> Thanks
Terry

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:205994
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: High Load Server... how much more can it take?

[Terry Ford]

Thanks for the responses.  Lots of good food for thought.

The reason that I initially went with database-bound client vars was due to an 
early CFMX bug whereby cookies weren't getting set correctly in all browsers.  
I believe it was fixed in CFMX 6.1, but since everything was smoothly running 
on the setup at the time didn't bother changing it then.   

I made the change back to client cookies today, and it appears to be working 
largely as intended.  

However, cookies still seem to behave erratically.  For example, I find a 
client var set (setdomaincookies=yes) from a page at "domain.com/login.cfm" is 
usually also found on a page at "www.domain.com/login.cfm" in IE,  but not in 
Firefox.  Firefox isn't acknowledging "www.domain.com" and "domain.com" as 
equivalent for purposes of client cookies.

To make things even more confusing, if I delete a client variable in 
www.domain.com/test.cfm then re-create it in www.domain.com/test2.cfm, 
sometimes IE doesn't even acknowledge that it has been set.  IE seems to get 
confused and will refuse to set the cookie correctly until it's either rebooted 
or you use "clear all cookies" in the browser.   

I honestly am not sure what the deal is, but I'm finding it very frustrating 
trying to get multiple browsers to react consistently to "cookie" client 
variables.  Right now it seems to me that client variables (type=cookie), when 
set, do one of three things in the same browser, depending on how the browser 
is feeling at the time:

- cli var gets set correctly, and is acknowledged under *.domain.com
- cli var gets set correctly, but is only acknowledged under current.domain.com
- doesn't get set at all. browser entirely ignores it.  It tends to exhibit 
this behavior if the client variable was just recently deleted and then 
re-established (such as a login after a logout).

These behaviors disappear the INSTANT I turn client storage from "cookie" to 
"database". 

If anyone has any insight into what the problem may be, please let me know.  As 
I'm on update 6.1, I didn't think I needed the "setdomaincookies" hot fix, 
which is listed on the MM site as an MX 6.0 fix.  Whose fault is this erratic 
behavior... CF's, IE's, or mine?

Thanks
Terry

~|
Discover CFTicket - The leading ColdFusion Help Desk and Trouble 
Ticket application

http://www.houseoffusion.com/banners/view.cfm?bannerid=48

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:205988
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

High Load Server... how much more can it take?

[Terry Ford]

Hi all,

I run a large entertainment website, using ColdFusion MX.  It has evolved over 
the past 9 yrs or so from tcl to php to CF4 to MX.

We serve over 12,000,000 .cfm pages a month right now with over 700,000 uniques 
per month. 

Our single server is a dual xeon 2.0 with 4 gigs of RAM. We use redhat, mysql 
and httpd.

Pretty simple setup really, and I'd like to keep it that way.  I am a big fan 
of optimization.

Unfortunately, the server has started bogging down in peak hours, and I think 
I'm nearing the end of the "optimize it" challenge. I've optimized queries, 
tweaked settings, query cached, cf_accelerated, sucked frequent db stuff into 
application scope, etc. However, it appears that there's simply too much load 
at peak hours to keep the site running fast.

Most of our templates are the usual efficent run-of-the-mill stuff, CFQUERYs 
and outputs, session and client variables (to db).  

cfstat shows approximately 20 pages/second during peak, where server load 
spikes up to 5 or 6, and requests start queueing causing nasty latencies.

A few questions:

1. Does 12,000,000 .cfm pages / month (distributed normally over the day, 
peaking around 4 PM ET) seem like a lot for a single xeon 2.0?

Do you think its possible for me to squeeze a lot more life out of this machine 
by investigating further optimizations, or does it sound like it's time to 
abandon that exercise and get a second machine? Can anyone else who is using a 
dual xeon on linux heavily let me know what sort of load / pages the thing is 
successfully serving? 

2. I'm not so sure that we're using RAM effectively, even with mysql set to use 
lots of memory and cf set to cache lots of queries and pages.   

"Free" shows the following stats.  It appears to me that a large amount of 
memory isn't being used by cf or mysql, even when the site starts slowing down:

Mem:   40680964023912  44184  0  555163321412
-/+ buffers/cache: 6469843421112
Swap:  2048276  347522013524

If I'm reading this correctly, there's a ton of memory doing nothing while the 
machine starts suffocating.  Any ideas on how I can put that memory to use? 

3.  Could the use of database client variables be causing peak hour 
sluggishness?  I've always wondered what sort of performance hit using database 
client variables has under load.  Would I get a noticable improvement switching 
to cookie client variables?

Thanks for your thoughts.

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:205940
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

CFMX and CGLOBAL/CDATA not purging

[Terry Ford]

Hey all,

In the process of performing a yearly upgrade I was quite surprised to find 
that my CGLOBAL and CDATA tables contained over 20 million (yes, million) 
entries each, and the CGLOBAL.MYD file alone had a size of 3.1 gigs.

Needless to say, I had been operating under the assumption that the tables were 
purging correctly.  Apparantly I was incorrect.

Anyway, I've taken a hatchet to the tabbles and whacked them down to size, and 
have verified that indeed the purge option (30 days) is selected.

Does anyone have any clue why my tables aren't purging? 
I am running mysql / CFMX6.1 / lotsa mem / dual xeon

Terry

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:192291
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Strange jrunserver.store errors

[Terry Ford]

Hi Dave,

Indeed, changing the permissions seem to have fixed half of the problem.  

Now my error log is getting filled with:

[Sat Jan 29 13:27:33 2005] [notice] jrApache[2460: 17675] initialized proxy for 
127.0.0.1:51010
[Sat Jan 29 13:27:34 2005] [notice] jrApache[2462: 42244] initialized proxy for 
127.0.0.1:51010
[Sat Jan 29 13:27:34 2005] [notice] jrApache[2463: 35160] initialized proxy for 
127.0.0.1:51010

Any idea why (a) it's doing this continually, and (b) why this is considered an 
error?  Is my proxy continually dying and having to be reset? 

Thanks

~|
Logware (www.logware.us): a new and convenient web-based time tracking 
application. Start tracking and documenting hours spent on a project or with a 
client with Logware today. Try it for free with a 15 day trial account.
http://www.houseoffusion.com/banners/view.cfm?bannerid=67

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:192192
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Strange jrunserver.store errors

[Terry Ford]

Hi all,

Running CFMX 6.1.  Have noticed that I'm getting a ton of these errors in my 
apache error_log:

[Fri Jan 28 21:51:16 2005] [notice] jrApache[11155: 64495] initialized proxy 
for 127.0.0.1:51010
[Fri Jan 28 21:51:16 2005] [notice] jrApache[11155: 64495] could not open 
"/opt/coldfusionmx/runtime/lib/wsconfig/1/jrunserver.store": Permission denied

These tend to appear repeated 3 or more times at the same second in the log, 
and occur tens or hundreds of times each hour.  

The website is under fairly heavy load.
If anyone has any idea what may be causing these or how to correct the problem, 
please let me know.

Thanks,
Terry

~|
Find out how CFTicket can increase your company's customer support 
efficiency by 100%
http://www.houseoffusion.com/banners/view.cfm?bannerid=49

Message: http://www.houseoffusion.com/lists.cfm/link=i:4:192154
Archives: http://www.houseoffusion.com/cf_lists/threads.cfm/4
Subscription: http://www.houseoffusion.com/lists.cfm/link=s:4
Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4
Donations & Support: http://www.houseoffusion.com/tiny.cfm/54

Re: Verity on Red Hat 9

[Terry Ford]

Has anyone here got Verity running on RH 9?

I'm using the most recent compat libraries, so I'm not
entirely sure what the next step is for trying to get Verity to
work.  

Terry

  - Original Message - 
  From: Thomas Chiverton 
  To: CF-Talk 
  Sent: Monday, December 01, 2003 11:55 AM
  Subject: Re: Verity on Red Hat 9

  On Monday 01 Dec 2003 13:46 pm, Yves Arsenault wrote:
  > I thought there were issues posted on the MM site about Verity and RedHat
  > 7.3 +.
  > To quote the site: " Verity does not support RedHat Linux 7.3, 8.0, and 9.

  "Does not support" is not the same as "Can't be made to run" as many people 
  will testify :-)

  -- 
  Tom Chiverton 
  Advanced ColdFusion Programmer

  Tel: +44(0)1749 834997
  email: [EMAIL PROTECTED]
  BlueFinger Limited
  Underwood Business Park
  Wookey Hole Road, WELLS. BA5 1AF
  Tel: +44 (0)1749 834900
  Fax: +44 (0)1749 834901
  web: www.bluefinger.com
  Company Reg No: 4209395 Registered Office: 2 Temple Back East, Temple
  Quay, BRISTOL. BS1 6EG.
  *** This E-mail contains confidential information for the addressee
  only. If you are not the intended recipient, please notify us
  immediately. You should not use, disclose, distribute or copy this
  communication if received in error. No binding contract will result from
  this e-mail until such time as a written document is signed on behalf of
  the company. BlueFinger Limited cannot accept responsibility for the
  completeness or accuracy of this message as it has been transmitted over
  public networks.***
 [Todays Threads] 
 [This Message] 
 [Subscription] 
 [Fast Unsubscribe] 
 [User Settings]

Verity on Red Hat 9

[Terry Ford]

Hey folks,

Just wondering if anyone has experience with Red Hat 9 Verity problems with
CF6.1.  Right now CFMX crashes within hours with a log file (hs_err_*) after
CFSEARCHes are executed.  Some of the searches work.  Some of them
cause a crash.  Haven't noticed any patterns yet.

Unexpected Signal : 11 occurred at PC=0x4037FAC7
Function=(null)+0x4037FAC7
Library=/opt/coldfusionmx/runtime/jre/jre/lib/i386/client/libjvm.so

Current Java thread:
at coldfusion.tagext.NativeCfx.processRequest(Native Method)
at coldfusion.tagext.lang.CFSearchTag.doStartTag(CFSearchTag.java:207)
at coldfusion.runtime.CfJspPage._emptyTag(CfJspPage.java:1871)
at cfsearch2_act2ecfm552951320.runPage(/home/trivia/www/search2_act.cfm:119)

It did not do this on Linux 7.2.  So it appears that CFSEARCH is just dying in a
RH 9 specific manner.  Any ideas on what I ought to consider patching or tweaking
to get Verity to work would be much appreciated.  

Thanks,
Terry

 [Todays Threads] 
 [This Message] 
 [Subscription] 
 [Fast Unsubscribe] 
 [User Settings]

Re: Internationalization Issue

[Terry Ford]

> terry, can you elaborate on this? i never use any of the LS functions (as
i
usually work w/locales that aren't supported) & have never seen this sort of
behavior.

Ok, here's the scenario:

I've got 2 templates.  Both query a db and return some accented French text
and a date.  CFOUTPUT displays the text.  Template 1 displays the date
using DateFormat.  Template 2 displays the date using LSDateFormat.

Upon execution template 1 does not display the French characters correctly.
They come out as boxes.  Template 2 does display the characters correctly.

I'm running linux 7.1 and JVM 1.4.2.  I'm running RedSky right now, but I
think I noticed this behavior a few weeks ago under Update 3.  I'll
investigate
to see if this is an outstanding bug.

Terry


- Original Message - 
From: "Paul Hastings" <[EMAIL PROTECTED]>
To: "CF-Talk" <[EMAIL PROTECTED]>
Sent: Wednesday, July 09, 2003 3:56 AM
Subject: Re: Internationalization Issue


> > After digging further, I found that it was the presence of a time
function
> on
> > the pages that didn't work that was prohibiting CF from outputting the
> > accents properly.  A simple change of DateFormat to LSDateFormat() fixed
> > the problem, and the accents magically started appearing.
>
> terry, can you elaborate on this? i never use any of the LS functions (as
i
> usually work w/locales that aren't supported) & have never seen this sort
of
> behavior.
>
>
> 
~|
Archives: http://www.houseoffusion.com/cf_lists/index.cfm?forumid=4
Subscription: 
http://www.houseoffusion.com/cf_lists/index.cfm?method=subscribe&forumid=4
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq

Get the mailserver that powers this list at 
http://www.coolfusion.com

Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

Re: Internationalization Issue

[Terry Ford]

Hey Richard,

I had the identical problem a couple of days ago.  It turned out that some
of my templates were showing accents, and others weren't, even with the
content declaration. 

After digging further, I found that it was the presence of a time function on
the pages that didn't work that was prohibiting CF from outputting the 
accents properly.  A simple change of DateFormat to LSDateFormat() fixed 
the problem, and the accents magically started appearing.

So take a look for any date/time functions in use on your internationalized
pages and make sure you've got the int'll versions of them.  Neo gets
cranky otherwise.

Terry

-Original Message- 
From: Richard Crawford [mailto:[EMAIL PROTECTED] 
Sent: 04 July, 2003 01:09 
To: CF-Talk 
Subject: Re: Internationalization Issue 

At 04:52 PM 7/3/2003 -0400, Gabriel Robichaud wrote: 

>I had a similar issue on win2k and added the following to my 
>Application.cfm file :  and it solved all my foreign euro-accents problems. 

Thanks for the suggestion. Unfortunately, it didn't work out. 

When submitting a form -- it's a quiz in our Spanish class -- I enter, 

--> son de Panamá 

When it gets served back to me, it reads: 

--> son de Panamá 

which the quiz marks as wrong. 

Richard 


~|
Archives: http://www.houseoffusion.com/cf_lists/index.cfm?forumid=4
Subscription: 
http://www.houseoffusion.com/cf_lists/index.cfm?method=subscribe&forumid=4
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq

This list and all House of Fusion resources hosted by CFHosting.com. The place for 
dependable ColdFusion Hosting.
http://www.cfhosting.com

Unsubscribe: 
http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=89.70.4

Re: Linux CF Error

[Terry Ford]

Hi Sarah,

I've got the identical setup: CF5, Apache 1.3.22-6, RH7.2, and the same
problem as you.

Approx twice a day I'm getting those messages in server.log:
server.log:"Fatal","TID=9225","07/15/02","16:15:19","Caught a fatal signal
(11) - Aborting"

No clue at all why this is happening, except that I suspect that it may have
to do with the connector and apache. There are a whole pile of threads on
the MM forums from people with this problem.

Another (possibly related?) problem that I'm running into (and you may
eventually) is as follows:

After a couple of hours running, I get httpd threads hanging after executing
one of a couple of templates. A typical apache status line for one of these
hung threads looks like this:

19242 0/53/352 W 0.10 4837 0 0.0 0.24 1.75 64.12.96.75 www.site.com POST
/template.cfm?n=533 HTTP/1.0

As you can see, this thread has a W status (W = sending reply), and it has
been in this state for 4837 seconds. Hung. ColdFusion under CFSTAT also
lists this thread as running, but never times it out. An strace -p PID on
the PID gives "read(8," as the last system call of the httpd thread.  For
some reason it seems to be deadlocked AFTER executing the template, or
something.

What happens is that these hung threads continue to build up until CF runs
out of threads and CF just sits there until either httpd or coldfusion is
restarted. These hangs occur at any time, although are particularly frequent
during peak traffic times and heavier load.

Have you experienced this httpd problem at all?

Regards,
Terry



- Original Message -
From: "Sarah Atkinson" <[EMAIL PROTECTED]>
To: "CF-Talk" <[EMAIL PROTECTED]>
Sent: Tuesday, July 16, 2002 6:08 PM
Subject: RE: Linux CF Error


> CF v5.0
> Apache 1.3.22-6
> Redhat 7.2
>
> Both Redhat and Apache patched up.
>
> Thanks,
> Sarah.
>
> -Original Message-
> From: Matt Liotta [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, 17 July 2002 10:59 AM
> To: CF-Talk
> Subject: RE: Linux CF Error
>
> You gotta give us some more information. What version of CF? What web
> server? What Linux version? Etc.
>
> Matt Liotta
> President & CEO
> Montara Software, Inc.
> http://www.montarasoftware.com/
> V: 415-577-8070
> F: 415-341-8906
> P: [EMAIL PROTECTED]
>
> > -Original Message-
> > From: Sarah Atkinson [mailto:[EMAIL PROTECTED]]
> > Sent: Tuesday, July 16, 2002 5:53 PM
> > To: CF-Talk
> > Subject: Linux CF Error
> >
> >
> > Hi All,
> >
> > We are experiencing some problems with one of our Linux CF servers.
> > Frequently the CF service restarts itself, having received the
> following
> > error:
> >
> > "Fatal","8200","07/15/02","17:13:52",,"Caught a fatal signal (11) -
> > Aborting"
> >
> > Any pointers on resolving this issue would be greatly appreciated. (we
> > have
> > a feeling its Java related)
> >
> > Thanks,
> > Sarah.
> >
> >
> > Sarah Atkinson
> > ColdFusion Programmer
> >
> > Lonely Planet Publications, Aust.
> > Telephone: +61 3 8379 8000
> > Fax: +61 3 8379 8111
> > [EMAIL PROTECTED] 
> >
> >
> >
>
> 
__
Your ad could be here. Monies from ads go to support these lists and provide more 
resources for the community. http://www.fusionauthority.com/ads.cfm
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

Re: How to override RequestTimeout's in CFMX?

[Terry Ford]

Thanks ray :)  It's Terry Ford here btw.  Sort of lurking.  Taking some time
to build up my website since I left MM.

You wouldnt happen to know much about CFMX on linux would you?  I'm
encountering a weird problem..

Terry

- Original Message -
From: "Raymond Camden" <[EMAIL PROTECTED]>
To: "CF-Talk" <[EMAIL PROTECTED]>
Sent: Tuesday, July 16, 2002 6:19 AM
Subject: RE: How to override RequestTimeout's in CFMX?


> You can set the value using a new attribute of cfsettting,
> requesttimeout.
>
> ===
> Raymond Camden, ColdFusion Jedi Master for Macromedia
>
> Email: [EMAIL PROTECTED]
> Yahoo IM : morpheus
>
> "My ally is the Force, and a powerful ally it is." - Yoda
>
> > -Original Message-
> > From: SoW [mailto:[EMAIL PROTECTED]]
> > Sent: Tuesday, July 16, 2002 10:07 AM
> > To: CF-Talk
> > Subject: How to override RequestTimeout's in CFMX?
> >
> >
> > CF5 allows for "?RequestTimeout=X" to be tacked on to any URL
> > to override the CFADMIN timeout.  Does anyone know if there
> > is any way to do this in CFMX?  It seems to ignore this param
> > as far as I can tell.
> >
> > I want 99.999% of my templates to have a timeout as a
> > protection, but be able to run some long-running templates
> > when specified.
> >
> > Thanks
> >
> >
> 
__
This list and all House of Fusion resources hosted by CFHosting.com. The place for 
dependable ColdFusion Hosting.
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

CF5/linux problem

[Terry Ford]

Hey folks,

I'm encountering a strange problem that maybe some of you have run into before.

Config: CF5, linux 7.2, Apache/CF5 EAPI module, mysql

I have one template that runs many times a minute.  However, under cfstat, a thread 
running this 
template seems to hang somewhat randomly -- every few hours or so. Under cfstat it is 
listed as a
running request.  What happens is that 10 such requests over time end up running and CF
grinds to a halt.  The template timeout of 45 seconds is ignored.  My current solution 
is to
cron-restart httpd every 12 hours -- not a pretty solution.

The strange part: the template does NOT seem to be caught in a loop.  If I restart 
Apache,
these infinitely-running threads disappear, and I get server.log entries:

"Warning","TID=11275","07/15/02","07:42:05","Template: /template.cfm, Ran: 5258 
seconds."
"Error","TID=8200","07/15/02","07:42:06","Unable to write reply -- client browser 
stopped waiting for request. "

If I create a new template with an intentional loop and run it, restarting Apache does 
not remove the thread
from the running list.

So, it SEEMS to me that CF is done with the template but that Apache is doing 
something weird.
Maybe?  

Does anyone have any experience or ideas?  I'm not entirely sure how to debug this one.

Thanks,
Terry


__
This list and all House of Fusion resources hosted by CFHosting.com. The place for 
dependable ColdFusion Hosting.
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

RE: To Beta or Not to Beta

[Terry Ford]

Actually, the criteria for beta 1 inclusion were fairly
basic.  We had a large number of applicants, and decided
to only invite a small number so that we could ensure that
we could better work through any initial problems with testers.

We took a look at all of the system configuration information
we had and selected from the pool to ensure a very diverse
beta 1 group that included developers of differing ability levels.  

This is certainly different to how we have run beta programs
in the past.  We have opted for a more structured approach
that should help us ensure that the Neo beta is as productive
for us and as pleasant an experience for testers as possible.

For those of you who were not included in beta 1 (the majority
of you), beta 2 is just around the corner, so never fear.

The Neo beta cycle is quite extensive for this release, so you 
will all have plenty of opportunity to give Neo a thorough workout.  

We're not limiting participants to be mean or evil or anything :)
We just want to make sure your first experience with Neo is a 
positive one.

Terry
Neo Product Management
Macromedia Server Products

-Original Message-
From: Jeffry Houser [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, February 06, 2002 11:48 AM
To: CF-Talk
Subject: To Beta or Not to Beta


At 03:35 AM 02/07/2002 +1100, you wrote:
>Jeffry I would agree, but I was told by MM that I would be notified
>when it goes to beta.

  I will admit that is a point I did not get in your original 
post.  ;)  However, who knows (Other than those wacky Macromedia-guys) 
how 
they chose people to be involved in Beta 1?  Presumably the Alpha was 
fairly limited (to select partners only I would presume).  So, perhaps 
the 
beta


>And that has to mean something by a development company, because if 
they
>can't go on that promise, how are we to trust anything they say.

  True, unfortunately, that is rarely the case.  ( Most glaringly 
exemplified by missed release dates)



--
Jeffry Houser | mailto:[EMAIL PROTECTED]
AIM: Reboog711  | ICQ: 5246969 | Fax / Phone: 860-223-7946
--
Need a Web Developer?  Contact me!
My Book: Instant ColdFusion 5  | http://www.instantcoldfusion.com
My New Book: ColdFusion: A Beginner's Guide February 2002
--
Far Cry Fly, Alternative Folk Rock
http://www.farcryfly.com | http://www.mp3.com/FarCryFly

__
Why Share?
  Dedicated Win 2000 Server · PIII 800 / 256 MB RAM / 40 GB HD / 20 GB MO/XFER
  Instant Activation · $99/Month · Free Setup
  http://www.pennyhost.com/redirect.cfm?adcode=coldfusionc
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

Neo Goes Beta!

[Terry Ford]

ColdFusion Community:

We are pleased to announce that the first beta of the next-generation of Macromedia 
ColdFusion Server, code-named Neo, was recently released to a small group of 
ColdFusion users who met our selection criteria. 

To ensure a great customer experience, the Neo Beta Program will include a series of 
beta releases. With each successive release, we plan to expand the group of 
participants to eventually include all ColdFusion customers who would like to 
participate. 

Apply now if you would like to participate in the next beta release of Neo at: 
http://www.macromedia.com/go/neobetaapply
Please apply only once.

The best way to ensure you are notified for future ColdFusion betas is to register 
your copy of ColdFusion Server or ColdFusion Studio at 
http://www.macromedia.com/software/register.

Regards,
Macromedia ColdFusion Beta Team

Terry Ford
ColdFusion Product Management
__
Get Your Own Dedicated Windows 2000 Server
  PIII 800 / 256 MB RAM / 40 GB HD / 20 GB MO/XFER
  Instant Activation · $99/Month · Free Setup
  http://www.pennyhost.com/redirect.cfm?adcode=coldfusionb
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists

Vote on ColdFusion Enhancements!

[Terry Ford]

Hi cf-talk,

We're deep in the middle of the design and development for the
next major release of ColdFusion and are looking for some good 
input on ColdFusion enhancements.  Below is a link to an 
enhancement-tracking application that we've put online temporarily.
Inside you will find over 200 CF enhancement requests that we have 
received and logged over the past year that you may vote on or add 
to.  

You will need to create a simple account to vote. Note that this 
system will only be available publicly for two days as we need to 
put this application on a solid server before opening up 
permanently -- we'd just like some quick feedback now.

http://syndication.allaire.com/vote/

Your detailed input is important.  Please post comments in the 
app beside any features or enhancements that interest you.  Not 
only is your vote important, but also descriptions about how you'd 
use such a feature or a story about how you needed such a feature 
in the past.  Please be as verbose as possible when posting -- 
more info, the better!  Feel free to not only vote on items but to 
also add your own items.

Standard disclaimer:

Other variables go into the blender of feature definition.  
100 votes for "I'd like CF to emulate the Vic-20" does not 
necessarily mean that this feature will make the cut.  This is 
simply one of a number of methods that we're using to help guide
product development.

Thanks,
Terry 
ColdFusion Product Management


~~
Structure your ColdFusion code with Fusebox. Get the official book at 
http://www.fusionauthority.com/bkinfo.cfm
FAQ: http://www.thenetprofits.co.uk/coldfusion/faq
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
Unsubscribe: http://www.houseoffusion.com/index.cfm?sidebar=lists