Adam,
I hear what you're saying (i.e. if a race condition is possible lock
it whether it matters or not as a matter of course) and I don't think
you're trying to give me a hard time. Just a thoughtful discussion.
With the statements at corfield.org that these guidelines are posted
by the Macrom
> > I don't try to second-guess unusual scenarios under which race
conditions
> > might not need locking; I just lock them according to the rules,
>
> C'mon, I asked a very specific question. I know my example is
> frivolous but it is oversimple for the sake of easy illustration.
And I promise th
is poorly understood by the CF community.
I don't buy that it would wrong at this late date.
-Original Message-
From: Adam Churvis [mailto:[EMAIL PROTECTED]
Sent: Monday, 9 May 2005 7:20
To: CF-Talk
Subject: Re: High Load Server... how much more can it take?
[snip]
> "...r
One last thing about that 'frivolous' example I mentioned above. The
one with the user count. I think that the reasons for locking in the
application scope -- other than the separate constants I also
mentioned -- are likely to be compelling and nearly universal, if not
completely so. A better ex
> I don't try to second-guess unusual scenarios under which race conditions
> might not need locking; I just lock them according to the rules,
C'mon, I asked a very specific question. I know my example is
frivolous but it is oversimple for the sake of easy illustration.
I gave a specific exampl
Sorry, Matt, I didn't answer your question:
> Now its my turn to say 'thats just plain wrong'. Or perhaps instead
> "tell my why there is any reason whatsoever that code which creates a
> benign race condition should be locked."
My reason is that, though the code is sacrificial, you should do yo
> On 5/8/05, Adam Churvis <[EMAIL PROTECTED]> wrote:
> > > Raymond Camden's "Tips for CFMX-ifying your ColdFusion 5
> > > Applications"
> >
(http://www.macromedia.com/devnet/mx/coldfusion/articles/updating_legacy.htm
> > l)
> > However a cflock will not help you in such a case.
> > > You'll get on
On 5/8/05, Adam Churvis <[EMAIL PROTECTED]> wrote:
> > Raymond Camden's "Tips for CFMX-ifying your ColdFusion 5
> > Applications"
> (http://www.macromedia.com/devnet/mx/coldfusion/articles/updating_legacy.htm
> l)
> However a cflock will not help you in such a case.
> > You'll get one set of write
> Raymond Camden's "Tips for CFMX-ifying your ColdFusion 5
> Applications"
(http://www.macromedia.com/devnet/mx/coldfusion/articles/updating_legacy.htm
l)
> discusses the need to only lock in case of race conditions. If your
> application uses *any* persistent scopes (client vars included) you
>
Oh and another thing. I use client vars extensively to maintain state
in my applications. Mostly out of an unfortunate need to maintain
ColdFusion 5 clustering compatibility. I have to say I can't wait to
be able o drop CF5 legacy support but those days are at least a year
away. Anyway, I have
Raymond Camden's "Tips for CFMX-ifying your ColdFusion 5
Applications"
(http://www.macromedia.com/devnet/mx/coldfusion/articles/updating_legacy.html)
discusses the need to only lock in case of race conditions. If your
application uses *any* persistent scopes (client vars included) you
have to de
That's right - when the exclusive lock tries to go into effect the read only
lock does something :P
Poor wording on my part.
-Original Message-
From: Adam Churvis [mailto:[EMAIL PROTECTED]
Sent: Sunday, 8 May 2005 9:52
To: CF-Talk
Subject: Re: High Load Server... how much more c
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
o
> Yes, this can't be stressed enough - if you don't use a read-only lock
where
> you read the data that is elsewhere exclusively locked, then your app will
> happily read data while it is being written to, giving you erroneous
results
> if it happens to occur at the right time.
Amen, brother.
> R
ks only do anything
while an exclusive lock is in effect so they don't affect performance
unnecessarily.
-Original Message-
From: Adam Churvis [mailto:[EMAIL PROTECTED]
Sent: Sunday, 8 May 2005 2:37
To: CF-Talk
Subject: Re: High Load Server... how much more can it take?
[snip]
There
> (Slightly OT) This is something I've wondered about for a while. It's
> obvious why you need to CFLOCK writes to the application scope since
> there are likely to be many users accessing application scope
> variables all over the place. I've heard that locking session
> variables is important i
> Adam is right on when he says its bad advice to consider cookies as a
> blanket substitute for client variables. They exist for good reasons.
> I can remember back to earlier versions of CF where they added
> significant stability to a CF application versus session variables.
Actually I was re
> Store your meaningless stuff in cookies. Only use them in cases where
> you will never need to set the cookie and read back from it within the
> same template. Never use them if you set them and cflocation to
> something else that needs to use the cookie value (both of these
> limitations can b
Terry,
Adam is right on when he says its bad advice to consider cookies as a
blanket substitute for client variables. They exist for good reasons.
I can remember back to earlier versions of CF where they added
significant stability to a CF application versus session variables.
However, CF has g
> If you've cached everything possible it might just be that your
application
> doesn't have enough data to use that much RAM. You might consider setting
> aside some of it as a RAM disk, but I'm not sure how well mySQL could use
> this.
>
> The idea would be to load your database (or the most use
> -Original Message-
> From: Terry Ford [mailto:[EMAIL PROTECTED]
> Sent: Saturday, May 07, 2005 1:26 AM
> To: CF-Talk
> Subject: High Load Server... how much more can it take?
>
>
> A few questions:
>
> 1. Does 12,000,000 .cfm pages / month (distributed no
You could move off Apache unless you really need anything it offers.
Is the machine swapping a lot (check out vmstat), or just waiting for queries
to come back ?
Turn off the client var updates (in the CFAdmin, properties for the client
store) and make sure purge is enabled (you are patched up t
> 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?
Not sure on this. Until you start coming close to maxing out the CPU
or memory, I'd say you should stick with your current hardware.
> 2. I'm not so sure
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
24 matches
Mail list logo
