Hi Guys,
Sorry this is turning into a long thread, but hopefully this will be
something that benefits others as well.
I've got Fusion Reactor running now, and I can notice quite a few
spikes up to 100% CPU looking at the graph on the System Metrics
page.
Also, memory usage is consistently up
The 216MB allocated should be the 'Heap' which is the memory the JVM gets to
use.
The 176MB is probably just the amount that is currently being used by the
JVM.
Could you paste in the JVM config line from the jvm.config file in Jrun /
bin / folder? that should say what the maximum allowable
In addition to Barry's helpful comments, I'll note something important about
FusionReactor's memory graph, as contrasted to SeeFusion or that of the CF
Server Monitor. As you clarified, Andrew, what you're referring to is the
heap amount allocated and used. That is not the max amount. Fortunately,
There are two things to keep in mind here:
1. Heap size and minimum/start heap size vs. maximum heap size.
I agree the best way to figure that out is to look at the jvm.config file.
From my experience it's _generally_ a good idea to set min = max
heap size right from the start, provided you
i thought this came up on a recent thread, but now I can't remember.
If you are getting lots of CPU spikes - you need to thread dump, thread dump
and then thread dump some more.
That way you can actually see what is going on on the server, and start to
look at what code is executing when.
I
cfjobs is kinda dead, so I'll post here
Hiring more ColdFusion developers in Melbourne
http://www.seek.com.au/job/coldfusion-developer/melbourne-inner/16693525/103
/1/
Regards
Dale Fraser
http://dale.fraser.id.au
http://learncf.com
http://flexcf.com
--
You received this
It's not dead, Dale, there's just no activity on CFJobs.People
are still watching it. Well I am anyway.
Cheers
Mike Kear
Windsor, NSW, Australia
Adobe Certified Advanced ColdFusion Developer
AFP Webworks
http://afpwebworks.com
ColdFusion, PHP, ASP, ASP.NET hosting from AUD$15/month
On
Is this what you guys all mean by the thread dumps? I got this from
Resources - List All Threads - Stack Trace All
Thread Stack TraceTrace Time: 10:46:20.870 02-Feb-2010Thread ID:
jms-fifo-1Priority:
Yup, that's the stuff.
You're looking for threads name jppr* (If I remember correctly), they will
show you what CF is doing.
Mark
On Tue, Feb 2, 2010 at 11:12 AM, Andrew Myers am2...@gmail.com wrote:
Is this what you guys all mean by the thread dumps? I got this from
Resources - List All
Hi,
Does anyone know if there is a way to generate a thread dump
automatically when server response times increase? I know how to do
one manually, but that's no good if I'm not around when the response
times increase. Does FusionReactor do this by any chance?
Thanks
Barry
--
You received this
Hi Barry,
I've just got hold of this
java.args=-server -Djava.awt.headless=true -Xmx1024m
-Dsun.io.useCanonCaches=false -XX:MaxPermSize=256m
-Dcoldfusion.rootDir={application.home}/
-Dcoldfusion.libPath={application.home}/lib -XX:+UseParallelGC
Hi People
Bit O.T. but anyone interested in telling me why the following method in
safari is so innccurate?
navigator.geolocation.getCurrentPosition()
yet the google maps application is so spot on?
I've even tried to recuse over that for a bit as suggested to little effect
thanks in advance
MaxPermSize set the PermGen size
Xmx is the Maximum heap size
Mark
On Tue, Feb 2, 2010 at 11:22 AM, Andrew Myers am2...@gmail.com wrote:
Hi Barry,
I've just got hold of this
java.args=-server -Djava.awt.headless=true -Xmx1024m
-Dsun.io.useCanonCaches=false -XX:MaxPermSize=256m
Also:
-Xms is mininum (start) heap size, I usually do:
-Xms1024m -Xmx1024 (in your case)
There's also -XX:PermSize for the minimum (start) permanent generation size.
Note: I wouldn't tweak the perm size without having a good case for doing so,
i.e. something that points to that a different
I think it's jrpp (JRun Proxy Port) - that'd be the case if you're using the
HTTP-Server
connector. IF you're using the internal HTTP server, the threads have a
different identifier
prefix iirc.
Cheers
Kai
Yup, that's the stuff.
You're looking for threads name jppr* (If I remember
Actually (was missing the m)
-Xms1024m -Xmx1024m
Cheers
Kai
Also:
-Xms is mininum (start) heap size, I usually do:
-Xms1024m -Xmx1024 (in your case)
There's also -XX:PermSize for the minimum (start) permanent generation size.
Note: I wouldn't tweak the perm size without
Not many watch it.
I have more success if I post here.
Just trying to get the message to as many CFer's as possible.
Regards
Dale Fraser
http://dale.fraser.id.au
http://learncf.com
http://flexcf.com
-Original Message-
From: cfaussie@googlegroups.com [mailto:cfaus...@googlegroups.com]
Thanks again Guys.
I've just watching the FR graphs and I noticed a big 100% spike at 11:54
which coincided with a long running request. I think I will investigate
this some more...
Andrew.
On 2 February 2010 11:52, Kai Koenig k...@koeni.de wrote:
I think it's jrpp (JRun Proxy Port) -
There is only about ~50 people difference b/w cfaussie and cfjobs
http://groups.google.com/group/cfaussie?hl=en
http://groups.google.com/group/cfjobs
So you aren't missing many eyeballs.
Mark
On Tue, Feb 2, 2010 at 12:07 PM, Dale Fraser d...@fraser.id.au wrote:
Not many watch it.
I have
136 actually.
Which is a large number of people, 1 of those 136 might want this job J
Regards
Dale Fraser
http://dale.fraser.id.au
http://learncf.com
http://flexcf.com
From: cfaussie@googlegroups.com [mailto:cfaus...@googlegroups.com] On Behalf
Of Mark Mandel
Sent: Tuesday, 2
Update:
This page generates a report from a query that itself takes about 29 seconds
(and I am not sure I can do a lot to improve that without getting some
cooperation from the guys who write the product I query out of, and that
isn't likely to happen quickly).
I've been caching the page with a
cfloop itself isn't bad.
You're just doing a lot of work in a single thread. I.e. you have a selfish
thread.
I guess the question is what is this doing in that thread that is so
important, and can it be optimised.
Is it the only place that causes the CPU to spike?
Mark
On Tue, Feb 2, 2010
Yeah, do what Kai said and set the minimum heap size to the same as the
maximum (means the heap does not have to keep resizing itself as it starts
filling up, heap resizing causes full garbage collections before the
resize).
The permanent generation size (PermSize) is actually additional to the
Additional special guest Steve Onnis :)
_
From: Paul Kukiel [mailto:kuki...@gmail.com]
Sent: Monday, 1 February 2010 4:41 PM
To: cfaussie@googlegroups.com
Subject: Re: [cfaussie] CF Drinks, Melbourne, Wed 3 Feb
I'll be there aswell.
Paul Kukiel
http://blog.kukiel.net
On Mon, Feb 1,
Have fun guys.
Please give me plenty of warning before the next one so that I can arrange
an office visit to Melb to coincide with it. :-)
On 2 February 2010 13:10, Steve Onnis st...@cfcentral.com.au wrote:
Additional special guest Steve Onnis :)
--
*From:*
Perhaps we should make it a regular event, in addition to the formal
CFUG night Andrew. We'll see how folk feel tomorrow evening. In any
case, I'll try to get more lead time out there in general.
Cheers Peter
On 2 February 2010 13:13, Andrew Myers am2...@gmail.com wrote:
Have fun guys.
Andrew, does that fair bit of nested cflooping show a CFQUERY buried
within that loop? Where it's taking data from the one query and using it to
do another? If so, that's a classic case where perhaps a join would solve
the problem. CFML developers often fall into a trap of having CF do work
that
Yes, Barry. That's what I was referring to late last week:
On the matter of FR's CP, let's clarify that there are two uses of the CP
feature. The first is just to notify you when conditions exceed the limits
you mention (like a request taking more than x seconds, more than x
requests running at
Awesome, thanks Charlie, I did not remember reading that (or I read it
in a rush) :).
Barry.
On Feb 2, 4:36 pm, charlie arehart charlie_li...@carehart.org
wrote:
Yes, Barry. That's what I was referring to late last week:
On the matter of FR's CP, let's clarify that there are two uses of the
If you're stumped for optimisation of a long running task, then having
a scheduled task is a great idea, but even when the scheduled task
runs, unless it's optimised it could still cause a performance hit.
An idea might be to put a 'thread wait ()' in your loop to allow other
things to run
For some more reading, there is an article I posted about managing selfish
threads in CF:
http://www.compoundtheory.com/?action=displayPostID=373
This is generally for situations where you can't optimise any further, or is
something that is just simply going to be expensive no matter what you do.
Barry Mark - I think this will be very useful.
I've got it down to 6 seconds (with the db query cached), but it's still
smashing the CPU.
Each iteration of the loop itself is negligible (cftimer is reporting
between 0-32ms). I think it's just that there's too many records. Each one
needs to
32 matches
Mail list logo
