jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Enric Jaen <enricj...@yahoo.es>
Subject Re: Why does Jmeter have a huge memory footprint?
Date Wed, 05 Dec 2012 08:31:38 GMT
Maybe you could add these params to the JVM to generate a .hprof file : 

-XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=<your log dir>


to be  analysed with a java profiler [1] what is eating the memory.

HTH
/Enric Jaen


[1] for example eclipse MAT plugin  http://www.eclipse.org/mat/



________________________________
 De: chaitanya bhatt <bhatt.chaitanya@gmail.com>
Para: JMeter Users List <user@jmeter.apache.org> 
Enviado: Miércoles 5 de diciembre de 2012 0:27
Asunto: Re: Why does Jmeter have a huge memory footprint?
 
Still working on it. Sorry Philippe and Smuel. I will try my best to log
the defect within the next two days.

Thanks
Chaitanya M Bhatt


On Tue, Dec 4, 2012 at 2:13 PM, Philippe Mouawad <philippe.mouawad@gmail.com
> wrote:

> Hello,
> Still no Bugzilla created.
>
> By the way what is the sizes of your responses , zipped and not zipped ?
>
> Thanks
> Regards
> Philippe
>
> On Fri, Nov 30, 2012 at 12:55 AM, chaitanya bhatt <
> bhatt.chaitanya@gmail.com
> > wrote:
>
> > Sure Philippe.
> >
> > Thanks a ton!
> > Chaitanya
> >
> > On Thu, Nov 29, 2012 at 2:25 PM, Philippe Mouawad <
> > philippe.mouawad@gmail.com> wrote:
> >
> > > Hello,
> > > I suggest you open a bugzilla (there is no one for that) and provide:
> > >
> > >    - jmeter.properties and user.properties file
> > >    - jmeter.log
> > >    - jmeter file you used
> > >    - Test Plan
> > >    - And as you seem to have evaluated the memory some screenshots of
> > >    consumptions you are talking about
> > >
> > >
> > > Regards
> > >
> > > Philippe
> > >
> > > On Thu, Nov 29, 2012 at 10:52 PM, chaitanya bhatt <
> > > bhatt.chaitanya@gmail.com
> > > > wrote:
> > >
> > > >  I am running the test via console/terminal.
> > > >  I am not saving transaction response times or responses to files.
> > > >  I am measuring response time directly from the Wily (servlet
> response
> > > > time).
> > > >  I have response assertions which checks the body of the HTTP
> response.
> > > >  The IF controller just evaluates a variable and checks if the value
> > > > exceeds number 2. Interpreter condition and Evaluate for all children
> > is
> > > > Unchecked.
> > > >  Basically, I have made sure that the test suit is as leane as
> > possible.
> > > >  I don't have any soft of non-html/non-json/non-xml HTTP resource
> > > requests.
> > > >
> > > > Thanks
> > > > Chaitanya
> > > >
> > > > On Thu, Nov 29, 2012 at 1:27 PM, Philippe Mouawad <
> > > > philippe.mouawad@gmail.com> wrote:
> > > >
> > > > > Hello,
> > > > >
> > > > > On Thu, Nov 29, 2012 at 10:12 PM, chaitanya bhatt <
> > > > > bhatt.chaitanya@gmail.com
> > > > > > wrote:
> > > > >
> > > > > > Thanks for replying Philippe!
> > > > > >
> > > > > > I noticed that the test cases you guys used to test the
> performance
> > > of
> > > > > > Jmeter is far to small than real world cases. Your test case
had
> > > JUST 3
> > > > > > HTTP samplers with just 2 N-V pairs in each sample.
> > > > > >
> > > > >
> > > > > Of course, this test plan must be simple for simple setup and it's
> > > more a
> > > > > basis to compare versions.
> > > > > It just illustrates the efforts made since 2.6 to improve
> > performances,
> > > > > look at dev list to see
> > > > > what was made on 2.8.
> > > > >
> > > > >
> > > > >
> > > > > > Real world scenarios have at least 100-200 HTTP samples per
> thread
> > > > group.
> > > > > > And the requests should contain at least 10-20 NV (name/value)
> > pairs
> > > > for
> > > > > > POST requests and the responses should at least weigh 80-120KB.
> At
> > > > least
> > > > > 20
> > > > > > response assertions. At least and at least 10 post response
regex
> > > > > > evaluators. This is when you actually exercise the engine.
> > > > > >
> > > > > > The test with 4000 threads I was talking about was a real life
> one.
> > > > >
> > > > > I am currently using Jmeter 2.8 and JRE 1.6.
> > > > > >
> > > > > > I am not using third party plugins. I am not using XML/CSV
> output.
> > I
> > > > have
> > > > > > removed all listeners.
> > > > > >
> > > > >
> > > > > I don't understand, which kind of save service are you using:
> > > > > CSV or  XML ?
> > > > >
> > > > > >
> > > > > > I have 250 HTTP samplers with at least 10-NV pairs in each
> request.
> > > > Most
> > > > > > requests are of type: POST. I have 35 assertions and 13 regex
> > > > > evaluators. I
> > > > > > have 2 If controllers, 15 transaction controllers, 8 constant
> > timers
> > > > and
> > > > > 1
> > > > > > thread group.
> > > > > >
> > > > >
> > > > > Which type of assertions ?
> > > > > What is the config of IfControllers ?
> > > > >
> > > > > >
> > > > > > Hope this helps. I will raise a defect in bugzilla but I'll
be
> > > > surprised
> > > > > if
> > > > > > it doesn't already have one.
> > > > > >
> > > > > > Yes please do and if possible attach a Test Plan as close to
> > possible
> > > > to
> > > > > your current one and anonymize data in it.
> > > > >
> > > > >
> > > > > > Thanks
> > > > > > Chaitanya M Bhatt
> > > > > >
> > > > > >
> > > > > >
> > > > > > On Thu, Nov 29, 2012 at 12:25 PM, Philippe Mouawad <
> > > > > > philippe.mouawad@gmail.com> wrote:
> > > > > >
> > > > > > > Hello,
> > > > > > > Which version of JMeter are you using ?
> > > > > > >
> > > > > > > Can you attach your test plan to a bugzilla or show a
> description
> > > of
> > > > > it ?
> > > > > > >
> > > > > > > As for what you are saying, I have already used jmeter
with a
> > > machine
> > > > > > > simular to the configuration you are describing and went
up to
> > 4000
> > > > > > threads
> > > > > > > and memory was not the limit in my case,
> > > > > > > so you should investigate your test plan first particularly
if
> > you
> > > > are
> > > > > > > talking about Perm Gen as it's not the kind of memory that
is
> > > highly
> > > > > > > consumed except if you are using Javascript for example.
> > > > > > >
> > > > > > >    - What kind of Test Elements are you using ?
> > > > > > >    - Are you using third party plugins?
> > > > > > >    - Which listeners have you setup
> > > > > > >    - Are you using XML or CSV output
> > > > > > >
> > > > > > >
> > > > > > > As for what you are saying regarding "*This is a frustrating
> > > problem
> > > > > > which
> > > > > > > has been overlooked for a long time.*", please see:
> > > > > > >
> > > > > > >    - http://wiki.apache.org/jmeter/JMeterPerformance
> > > > > > >
> > > > > > > Regards
> > > > > > >
> > > > > > > Philippe
> > > > > > >
> > > > > > > On Thu, Nov 29, 2012 at 8:40 PM, chaitanya bhatt
> > > > > > > <bhatt.chaitanya@gmail.com>wrote:
> > > > > > >
> > > > > > > > Group,
> > > > > > > >
> > > > > > > > I have noticed that Jmeter uses unreasonable amount
of memory
> > per
> > > > > > thread.
> > > > > > > > Even if you strip off loggers/result tree etc. you
would
> still
> > > see
> > > > a
> > > > > > huge
> > > > > > > > consumption of memory. I monitored the Jmeter memory
heap and
> > > tuned
> > > > > the
> > > > > > > JVM
> > > > > > > > as much as possible. In spite of this I noticed that
the
> > tenured
> > > > > > > generation
> > > > > > > > of the heap is always packed with objects of huge
size. Lots
> of
> > > > > classes
> > > > > > > are
> > > > > > > > loaded dynamically which is causing high occupancy
in perm
> gen.
> > > > > > > >
> > > > > > > > I am working on another home grown tool which uses
HTTPClient
> > > > library
> > > > > > to
> > > > > > > > generate load. Using this home grown tool I can generate
4000
> > > > > instances
> > > > > > > per
> > > > > > > > machine (64 bit with 32Gb ram and 16 CPU cores). But
Jmeter
> on
> > > the
> > > > > > > contrary
> > > > > > > > fails to scale to a mere 1/10th of the target user
load.
> > > > > > > >
> > > > > > > > This is a frustrating problem which has been overlooked
for a
> > > long
> > > > > > time.
> > > > > > > >
> > > > > > > > Can something be done about this?
> > > > > > > >
> > > > > > > > Thanks
> > > > > > > > Chaitanya M Bhatt
> > > > > > > > http://performancecompetence.com
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Cordialement.
> > > > > > > Philippe Mouawad.
> > > > > > >
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Cordialement.
> > > > > Philippe Mouawad.
> > > > >
> > > >
> > >
> > >
> > >
> > > --
> > > Cordialement.
> > > Philippe Mouawad.
> > >
> >
>
>
>
> --
> Cordialement.
> Philippe Mouawad.
>
Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message