Re: Graph Results overlap

2009-05-24 Thread Shadar

Hi everyone,

I am new to JMeter, and just ran into the same issue of the Graph Results
listener wrapping around (actually I posted a question to StackOverflow 
http://stackoverflow.com/questions/903348/jmeter-saving-results-configuring-graph-results-time-span
here  and then found this great group!). I don't know what the initial
rationale was and disagreeing with it would be a philosophical hence
pointless debate, but practically this is a big problem IMHO. I just ran a
weekend-long test, and I can only see the lines for the last 2 hours (just
guessing - there are no sample numbers or time labels on the X axis!)!

It seems to me that the simplest pragmatic solution would be to let user
change dynamically the time span (width) of the graph (just like I can check
boxes to add or remove lines), and then the graph should be repainted
according to the new width. This seems like a simple addition to the GUI and
logic. Any chance someone can do this?

A related question is if I can save the results of test runs and later on
open them with this listener? I know I can save the test plan or parts of
it. I am unclear if I can save separately just the test results data, and
later open them and perform comparisons etc. And furthermore can I open them
with different listeners even if they weren't part of original test (i.e. I
think of the test as accumulating data, and later on I want to view and
interpret the data using different viewers).

Thanks!

-- Shaul

P.S. To those familiar with MS's Performance Monitor (Perfmon), changing the
time span is one of the options when opening a saved performance log (data).
Other less important options IMHO are to change the line width, color and
type (e.g. dashed or dotted).
-- 
View this message in context: 
http://www.nabble.com/Graph-Results-overlap-tp23477206p23691823.html
Sent from the JMeter - User mailing list archive at Nabble.com.


-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



AW: [SPAM] Re: Graph Results overlap

2009-05-24 Thread B.Ramann
Hi Saul,

i have made good experience with j-meter running in gui mode and using the plug 
in stataggvisualizer. With this tool you can see the response time for each 
iteration (not per request) in a graph. As load generater is use remoter 
servers. When you use more then one remote-server, you can add or stop threads 
groups during the test and you can see also the response times in this graph. 
For the remote server instances, make sure that the stataggvisualizer libs are 
also installed.

Another good effect is when you using more remoter server instances, that you 
minimizes the out of mem problem.


Cheers
bjoern



 
-Ursprüngliche Nachricht-
Von: Shadar [mailto:shaul...@gmail.com] 
Gesendet: Sonntag, 24. Mai 2009 11:12
An: jmeter-user@jakarta.apache.org
Betreff: [SPAM] Re: Graph Results overlap
Wichtigkeit: Niedrig


Hi everyone,

I am new to JMeter, and just ran into the same issue of the Graph Results
listener wrapping around (actually I posted a question to StackOverflow 
http://stackoverflow.com/questions/903348/jmeter-saving-results-configuring-graph-results-time-span
here  and then found this great group!). I don't know what the initial
rationale was and disagreeing with it would be a philosophical hence
pointless debate, but practically this is a big problem IMHO. I just ran a
weekend-long test, and I can only see the lines for the last 2 hours (just
guessing - there are no sample numbers or time labels on the X axis!)!

It seems to me that the simplest pragmatic solution would be to let user
change dynamically the time span (width) of the graph (just like I can check
boxes to add or remove lines), and then the graph should be repainted
according to the new width. This seems like a simple addition to the GUI and
logic. Any chance someone can do this?

A related question is if I can save the results of test runs and later on
open them with this listener? I know I can save the test plan or parts of
it. I am unclear if I can save separately just the test results data, and
later open them and perform comparisons etc. And furthermore can I open them
with different listeners even if they weren't part of original test (i.e. I
think of the test as accumulating data, and later on I want to view and
interpret the data using different viewers).

Thanks!

-- Shaul

P.S. To those familiar with MS's Performance Monitor (Perfmon), changing the
time span is one of the options when opening a saved performance log (data).
Other less important options IMHO are to change the line width, color and
type (e.g. dashed or dotted).
-- 
View this message in context: 
http://www.nabble.com/Graph-Results-overlap-tp23477206p23691823.html
Sent from the JMeter - User mailing list archive at Nabble.com.


-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org


-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



Re: Graph Results overlap

2009-05-14 Thread Peter Lin
that might be a viable solution that is low cost, without writing an
auto-scaling graph.

peter

On Wed, May 13, 2009 at 9:48 PM, Ashley Martens amart...@ngmoco.com wrote:
 At the very least the graph should clear the old data when it starts back at
 the beginning.

 On May 13, 2009, at 5:45 PM, drubix wrote:


 I agree with you there.  I actually checked the documentation for any
 information regarding this odd behaviour and couldn't find any.  I do see
 where you're coming from though.

 All we really need is some sort of self-scaling graph so that instead of
 continually extending itself, it just redraws itself every now and then in
 a
 smaller size.

 I'll probably need to write some sort of graphing plugin so that I can
 include the results of the test in my final report for work anyway.  If I
 do
 end up doing this, I'll release the graphing plugin so that other people
 can
 use it too.

 Thanks for the heads up,

 Drew


 Peter Lin wrote:

 there's actually a good reason for it. It's not so much a design flaw
 as it is poor documentation.

 In order to not repeat, the graph could grow infinitely large,
 especially for long running tests. That would quickly chew up a ton of
 memory and make JMeter crash.  It was never meant to be a normal time
 series graph and was only meant to give a rough idea of what's going
 on.

 I'll give you an example. If I have a long running test that goes for
 3 days, how wide would the graph be? I've run tests for 1 week in the
 past with jmeter. What you're looking for is a different kind of
 graph, which currently doesn't exist in jmeter.  I thought about
 writing one in the past, but never got around to it.

 I haven't been active with jmeter since 2007. If you want a real time
 series graph, i would suggest implementing it yourself. You can look
 at the existing listeners to see how they work.

 In my mind, the bug is in the documentation, since it never really
 explains the purpose of the graph and the rationale behind it.

 peter


 On Wed, May 13, 2009 at 8:03 PM, drubix andrew.schr...@gmail.com wrote:

 Even if it's not a bug, it's definitely a design flaw.  The graph's
 produced
 don't make sense and can't be directly exported into a report due to
 their
 non-standard nature.  I'll file a bugzilla report regarding this.

 Drew


 Peter Lin wrote:

 I could be wrong, but it has always been that way. The default graph
 listener has a fixed width, so once it reaches the end it just
 continues at the beginning again.

 peter

 On Tue, May 12, 2009 at 8:29 PM, drubix andrew.schr...@gmail.com
 wrote:

 I've had this problem both when running the tests in GUI mode and
 non-GUI
 mode (and opening the JTL file after the test has completed).  I'm
 away
 from
 work today and don't have any of my testing files but I'll upload a
 bugzilla
 report tomorrow if that is not the intended functionality.

 Drew


 Peter Lin wrote:

 Isn't that how it has always worked. If the test plan is large and
 runs for a long time, the graph start back at the beginning again.

 peter

 On Tue, May 12, 2009 at 6:24 PM, sebb seb...@gmail.com wrote:

 On 11/05/2009, drubix andrew.schr...@gmail.com wrote:

  Hi,

  When using the Graph Results listener, if the plot goes off the
 end
 of
 the X
  axis it continues to plot at time 0 again.  This means that there
 are
 2
 or
  more lines running over the top of each other, all of the same
 colour.
  This
  produces some pretty nasty graphs which are difficult to read and
 are
  nonsensical.  I've attached an example:

  http://www.nabble.com/file/p23477206/graph.jpg

  I assume this isn't supposed to happen... does anybody know why
 mine
 is
  doing it?

 Not sure why this is happening; may be a threading issue.

 I suggest you save the samples to a JTL file - CSV mode is cheapest
 -
 and reload in the Listener after the test run. [This will be done in
 a
 single thread.]

 If it still causes wrap-round, then please raise a Bugzilla issue
 and
 attach the JTL file.

  Thanks,

  Drew

 --
  View this message in context:

 http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
  Sent from the JMeter - User mailing list archive at Nabble.com.



  -
  To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
  For additional commands, e-mail:
 jmeter-user-h...@jakarta.apache.org




 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org




 --
 View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23513354.html
 Sent from the JMeter - User mailing list archive

Re: Graph Results overlap

2009-05-13 Thread drubix

Even if it's not a bug, it's definitely a design flaw.  The graph's produced
don't make sense and can't be directly exported into a report due to their
non-standard nature.  I'll file a bugzilla report regarding this.

Drew


Peter Lin wrote:
 
 I could be wrong, but it has always been that way. The default graph
 listener has a fixed width, so once it reaches the end it just
 continues at the beginning again.
 
 peter
 
 On Tue, May 12, 2009 at 8:29 PM, drubix andrew.schr...@gmail.com wrote:

 I've had this problem both when running the tests in GUI mode and non-GUI
 mode (and opening the JTL file after the test has completed).  I'm away
 from
 work today and don't have any of my testing files but I'll upload a
 bugzilla
 report tomorrow if that is not the intended functionality.

 Drew


 Peter Lin wrote:

 Isn't that how it has always worked. If the test plan is large and
 runs for a long time, the graph start back at the beginning again.

 peter

 On Tue, May 12, 2009 at 6:24 PM, sebb seb...@gmail.com wrote:
 On 11/05/2009, drubix andrew.schr...@gmail.com wrote:

  Hi,

  When using the Graph Results listener, if the plot goes off the end
 of
 the X
  axis it continues to plot at time 0 again.  This means that there are
 2
 or
  more lines running over the top of each other, all of the same
 colour.
  This
  produces some pretty nasty graphs which are difficult to read and are
  nonsensical.  I've attached an example:

  http://www.nabble.com/file/p23477206/graph.jpg

  I assume this isn't supposed to happen... does anybody know why mine
 is
  doing it?

 Not sure why this is happening; may be a threading issue.

 I suggest you save the samples to a JTL file - CSV mode is cheapest -
 and reload in the Listener after the test run. [This will be done in a
 single thread.]

 If it still causes wrap-round, then please raise a Bugzilla issue and
 attach the JTL file.

  Thanks,

  Drew

 --
  View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
  Sent from the JMeter - User mailing list archive at Nabble.com.


  -
  To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
  For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org




 --
 View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23513354.html
 Sent from the JMeter - User mailing list archive at Nabble.com.


 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org


 
 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org
 
 
 

-- 
View this message in context: 
http://www.nabble.com/Graph-Results-overlap-tp23477206p23532109.html
Sent from the JMeter - User mailing list archive at Nabble.com.


-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



Re: Graph Results overlap

2009-05-13 Thread Peter Lin
there's actually a good reason for it. It's not so much a design flaw
as it is poor documentation.

In order to not repeat, the graph could grow infinitely large,
especially for long running tests. That would quickly chew up a ton of
memory and make JMeter crash.  It was never meant to be a normal time
series graph and was only meant to give a rough idea of what's going
on.

I'll give you an example. If I have a long running test that goes for
3 days, how wide would the graph be? I've run tests for 1 week in the
past with jmeter. What you're looking for is a different kind of
graph, which currently doesn't exist in jmeter.  I thought about
writing one in the past, but never got around to it.

I haven't been active with jmeter since 2007. If you want a real time
series graph, i would suggest implementing it yourself. You can look
at the existing listeners to see how they work.

In my mind, the bug is in the documentation, since it never really
explains the purpose of the graph and the rationale behind it.

peter


On Wed, May 13, 2009 at 8:03 PM, drubix andrew.schr...@gmail.com wrote:

 Even if it's not a bug, it's definitely a design flaw.  The graph's produced
 don't make sense and can't be directly exported into a report due to their
 non-standard nature.  I'll file a bugzilla report regarding this.

 Drew


 Peter Lin wrote:

 I could be wrong, but it has always been that way. The default graph
 listener has a fixed width, so once it reaches the end it just
 continues at the beginning again.

 peter

 On Tue, May 12, 2009 at 8:29 PM, drubix andrew.schr...@gmail.com wrote:

 I've had this problem both when running the tests in GUI mode and non-GUI
 mode (and opening the JTL file after the test has completed).  I'm away
 from
 work today and don't have any of my testing files but I'll upload a
 bugzilla
 report tomorrow if that is not the intended functionality.

 Drew


 Peter Lin wrote:

 Isn't that how it has always worked. If the test plan is large and
 runs for a long time, the graph start back at the beginning again.

 peter

 On Tue, May 12, 2009 at 6:24 PM, sebb seb...@gmail.com wrote:
 On 11/05/2009, drubix andrew.schr...@gmail.com wrote:

  Hi,

  When using the Graph Results listener, if the plot goes off the end
 of
 the X
  axis it continues to plot at time 0 again.  This means that there are
 2
 or
  more lines running over the top of each other, all of the same
 colour.
  This
  produces some pretty nasty graphs which are difficult to read and are
  nonsensical.  I've attached an example:

  http://www.nabble.com/file/p23477206/graph.jpg

  I assume this isn't supposed to happen... does anybody know why mine
 is
  doing it?

 Not sure why this is happening; may be a threading issue.

 I suggest you save the samples to a JTL file - CSV mode is cheapest -
 and reload in the Listener after the test run. [This will be done in a
 single thread.]

 If it still causes wrap-round, then please raise a Bugzilla issue and
 attach the JTL file.

  Thanks,

  Drew

 --
  View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
  Sent from the JMeter - User mailing list archive at Nabble.com.


  -
  To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
  For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org




 --
 View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23513354.html
 Sent from the JMeter - User mailing list archive at Nabble.com.


 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org




 --
 View this message in context: 
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23532109.html
 Sent from the JMeter - User mailing list archive at Nabble.com.


 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter

Re: Graph Results overlap

2009-05-13 Thread drubix

I agree with you there.  I actually checked the documentation for any
information regarding this odd behaviour and couldn't find any.  I do see
where you're coming from though.

All we really need is some sort of self-scaling graph so that instead of
continually extending itself, it just redraws itself every now and then in a
smaller size.

I'll probably need to write some sort of graphing plugin so that I can
include the results of the test in my final report for work anyway.  If I do
end up doing this, I'll release the graphing plugin so that other people can
use it too.

Thanks for the heads up,

Drew


Peter Lin wrote:
 
 there's actually a good reason for it. It's not so much a design flaw
 as it is poor documentation.
 
 In order to not repeat, the graph could grow infinitely large,
 especially for long running tests. That would quickly chew up a ton of
 memory and make JMeter crash.  It was never meant to be a normal time
 series graph and was only meant to give a rough idea of what's going
 on.
 
 I'll give you an example. If I have a long running test that goes for
 3 days, how wide would the graph be? I've run tests for 1 week in the
 past with jmeter. What you're looking for is a different kind of
 graph, which currently doesn't exist in jmeter.  I thought about
 writing one in the past, but never got around to it.
 
 I haven't been active with jmeter since 2007. If you want a real time
 series graph, i would suggest implementing it yourself. You can look
 at the existing listeners to see how they work.
 
 In my mind, the bug is in the documentation, since it never really
 explains the purpose of the graph and the rationale behind it.
 
 peter
 
 
 On Wed, May 13, 2009 at 8:03 PM, drubix andrew.schr...@gmail.com wrote:

 Even if it's not a bug, it's definitely a design flaw.  The graph's
 produced
 don't make sense and can't be directly exported into a report due to
 their
 non-standard nature.  I'll file a bugzilla report regarding this.

 Drew


 Peter Lin wrote:

 I could be wrong, but it has always been that way. The default graph
 listener has a fixed width, so once it reaches the end it just
 continues at the beginning again.

 peter

 On Tue, May 12, 2009 at 8:29 PM, drubix andrew.schr...@gmail.com
 wrote:

 I've had this problem both when running the tests in GUI mode and
 non-GUI
 mode (and opening the JTL file after the test has completed).  I'm away
 from
 work today and don't have any of my testing files but I'll upload a
 bugzilla
 report tomorrow if that is not the intended functionality.

 Drew


 Peter Lin wrote:

 Isn't that how it has always worked. If the test plan is large and
 runs for a long time, the graph start back at the beginning again.

 peter

 On Tue, May 12, 2009 at 6:24 PM, sebb seb...@gmail.com wrote:
 On 11/05/2009, drubix andrew.schr...@gmail.com wrote:

  Hi,

  When using the Graph Results listener, if the plot goes off the end
 of
 the X
  axis it continues to plot at time 0 again.  This means that there
 are
 2
 or
  more lines running over the top of each other, all of the same
 colour.
  This
  produces some pretty nasty graphs which are difficult to read and
 are
  nonsensical.  I've attached an example:

  http://www.nabble.com/file/p23477206/graph.jpg

  I assume this isn't supposed to happen... does anybody know why
 mine
 is
  doing it?

 Not sure why this is happening; may be a threading issue.

 I suggest you save the samples to a JTL file - CSV mode is cheapest -
 and reload in the Listener after the test run. [This will be done in
 a
 single thread.]

 If it still causes wrap-round, then please raise a Bugzilla issue and
 attach the JTL file.

  Thanks,

  Drew

 --
  View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
  Sent from the JMeter - User mailing list archive at Nabble.com.


  -
  To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
  For additional commands, e-mail:
 jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org




 --
 View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23513354.html
 Sent from the JMeter - User mailing list archive at Nabble.com.


 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user

Re: Graph Results overlap

2009-05-13 Thread Peter Lin
the documentation isn't as good as it should be.

a self scaling graph would be ideal.. It's just that no one got around
to implementing one.

peter

On Wed, May 13, 2009 at 8:45 PM, drubix andrew.schr...@gmail.com wrote:

 I agree with you there.  I actually checked the documentation for any
 information regarding this odd behaviour and couldn't find any.  I do see
 where you're coming from though.

 All we really need is some sort of self-scaling graph so that instead of
 continually extending itself, it just redraws itself every now and then in a
 smaller size.

 I'll probably need to write some sort of graphing plugin so that I can
 include the results of the test in my final report for work anyway.  If I do
 end up doing this, I'll release the graphing plugin so that other people can
 use it too.

 Thanks for the heads up,

 Drew


 Peter Lin wrote:

 there's actually a good reason for it. It's not so much a design flaw
 as it is poor documentation.

 In order to not repeat, the graph could grow infinitely large,
 especially for long running tests. That would quickly chew up a ton of
 memory and make JMeter crash.  It was never meant to be a normal time
 series graph and was only meant to give a rough idea of what's going
 on.

 I'll give you an example. If I have a long running test that goes for
 3 days, how wide would the graph be? I've run tests for 1 week in the
 past with jmeter. What you're looking for is a different kind of
 graph, which currently doesn't exist in jmeter.  I thought about
 writing one in the past, but never got around to it.

 I haven't been active with jmeter since 2007. If you want a real time
 series graph, i would suggest implementing it yourself. You can look
 at the existing listeners to see how they work.

 In my mind, the bug is in the documentation, since it never really
 explains the purpose of the graph and the rationale behind it.

 peter


 On Wed, May 13, 2009 at 8:03 PM, drubix andrew.schr...@gmail.com wrote:

 Even if it's not a bug, it's definitely a design flaw.  The graph's
 produced
 don't make sense and can't be directly exported into a report due to
 their
 non-standard nature.  I'll file a bugzilla report regarding this.

 Drew


 Peter Lin wrote:

 I could be wrong, but it has always been that way. The default graph
 listener has a fixed width, so once it reaches the end it just
 continues at the beginning again.

 peter

 On Tue, May 12, 2009 at 8:29 PM, drubix andrew.schr...@gmail.com
 wrote:

 I've had this problem both when running the tests in GUI mode and
 non-GUI
 mode (and opening the JTL file after the test has completed).  I'm away
 from
 work today and don't have any of my testing files but I'll upload a
 bugzilla
 report tomorrow if that is not the intended functionality.

 Drew


 Peter Lin wrote:

 Isn't that how it has always worked. If the test plan is large and
 runs for a long time, the graph start back at the beginning again.

 peter

 On Tue, May 12, 2009 at 6:24 PM, sebb seb...@gmail.com wrote:
 On 11/05/2009, drubix andrew.schr...@gmail.com wrote:

  Hi,

  When using the Graph Results listener, if the plot goes off the end
 of
 the X
  axis it continues to plot at time 0 again.  This means that there
 are
 2
 or
  more lines running over the top of each other, all of the same
 colour.
  This
  produces some pretty nasty graphs which are difficult to read and
 are
  nonsensical.  I've attached an example:

  http://www.nabble.com/file/p23477206/graph.jpg

  I assume this isn't supposed to happen... does anybody know why
 mine
 is
  doing it?

 Not sure why this is happening; may be a threading issue.

 I suggest you save the samples to a JTL file - CSV mode is cheapest -
 and reload in the Listener after the test run. [This will be done in
 a
 single thread.]

 If it still causes wrap-round, then please raise a Bugzilla issue and
 attach the JTL file.

  Thanks,

  Drew

 --
  View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
  Sent from the JMeter - User mailing list archive at Nabble.com.


  -
  To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
  For additional commands, e-mail:
 jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org




 --
 View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23513354.html
 Sent from the JMeter - User mailing list archive at Nabble.com.


 -
 To unsubscribe, e

Re: Graph Results overlap

2009-05-13 Thread Ashley Martens
At the very least the graph should clear the old data when it starts  
back at the beginning.


On May 13, 2009, at 5:45 PM, drubix wrote:



I agree with you there.  I actually checked the documentation for any
information regarding this odd behaviour and couldn't find any.  I  
do see

where you're coming from though.

All we really need is some sort of self-scaling graph so that  
instead of
continually extending itself, it just redraws itself every now and  
then in a

smaller size.

I'll probably need to write some sort of graphing plugin so that I can
include the results of the test in my final report for work anyway.   
If I do
end up doing this, I'll release the graphing plugin so that other  
people can

use it too.

Thanks for the heads up,

Drew


Peter Lin wrote:


there's actually a good reason for it. It's not so much a design flaw
as it is poor documentation.

In order to not repeat, the graph could grow infinitely large,
especially for long running tests. That would quickly chew up a ton  
of

memory and make JMeter crash.  It was never meant to be a normal time
series graph and was only meant to give a rough idea of what's going
on.

I'll give you an example. If I have a long running test that goes for
3 days, how wide would the graph be? I've run tests for 1 week in the
past with jmeter. What you're looking for is a different kind of
graph, which currently doesn't exist in jmeter.  I thought about
writing one in the past, but never got around to it.

I haven't been active with jmeter since 2007. If you want a real time
series graph, i would suggest implementing it yourself. You can look
at the existing listeners to see how they work.

In my mind, the bug is in the documentation, since it never really
explains the purpose of the graph and the rationale behind it.

peter


On Wed, May 13, 2009 at 8:03 PM, drubix andrew.schr...@gmail.com  
wrote:


Even if it's not a bug, it's definitely a design flaw.  The graph's
produced
don't make sense and can't be directly exported into a report due to
their
non-standard nature.  I'll file a bugzilla report regarding this.

Drew


Peter Lin wrote:


I could be wrong, but it has always been that way. The default  
graph

listener has a fixed width, so once it reaches the end it just
continues at the beginning again.

peter

On Tue, May 12, 2009 at 8:29 PM, drubix andrew.schr...@gmail.com
wrote:


I've had this problem both when running the tests in GUI mode and
non-GUI
mode (and opening the JTL file after the test has completed).   
I'm away

from
work today and don't have any of my testing files but I'll  
upload a

bugzilla
report tomorrow if that is not the intended functionality.

Drew


Peter Lin wrote:


Isn't that how it has always worked. If the test plan is large  
and
runs for a long time, the graph start back at the beginning  
again.


peter

On Tue, May 12, 2009 at 6:24 PM, sebb seb...@gmail.com wrote:

On 11/05/2009, drubix andrew.schr...@gmail.com wrote:


 Hi,

 When using the Graph Results listener, if the plot goes off  
the end

of
the X
 axis it continues to plot at time 0 again.  This means that  
there

are
2
or
 more lines running over the top of each other, all of the same
colour.
 This
 produces some pretty nasty graphs which are difficult to  
read and

are
 nonsensical.  I've attached an example:

 http://www.nabble.com/file/p23477206/graph.jpg

 I assume this isn't supposed to happen... does anybody know  
why

mine
is
 doing it?


Not sure why this is happening; may be a threading issue.

I suggest you save the samples to a JTL file - CSV mode is  
cheapest -
and reload in the Listener after the test run. [This will be  
done in

a
single thread.]

If it still causes wrap-round, then please raise a Bugzilla  
issue and

attach the JTL file.


 Thanks,

 Drew

--
 View this message in context:
http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
 Sent from the JMeter - User mailing list archive at  
Nabble.com.



  
-

 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail:
jmeter-user-h...@jakarta.apache.org




-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org




-
To unsubscribe, e-mail: jmeter-user- 
unsubscr...@jakarta.apache.org

For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org





--
View this message in context:
http://www.nabble.com/Graph-Results-overlap-tp23477206p23513354.html
Sent from the JMeter - User mailing list archive at Nabble.com.


-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org

Re: Graph Results overlap

2009-05-12 Thread sebb
On 11/05/2009, drubix andrew.schr...@gmail.com wrote:

  Hi,

  When using the Graph Results listener, if the plot goes off the end of the X
  axis it continues to plot at time 0 again.  This means that there are 2 or
  more lines running over the top of each other, all of the same colour.  This
  produces some pretty nasty graphs which are difficult to read and are
  nonsensical.  I've attached an example:

  http://www.nabble.com/file/p23477206/graph.jpg

  I assume this isn't supposed to happen... does anybody know why mine is
  doing it?

Not sure why this is happening; may be a threading issue.

I suggest you save the samples to a JTL file - CSV mode is cheapest -
and reload in the Listener after the test run. [This will be done in a
single thread.]

If it still causes wrap-round, then please raise a Bugzilla issue and
attach the JTL file.

  Thanks,

  Drew

 --
  View this message in context: 
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
  Sent from the JMeter - User mailing list archive at Nabble.com.


  -
  To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
  For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



Re: Graph Results overlap

2009-05-12 Thread Peter Lin
Isn't that how it has always worked. If the test plan is large and
runs for a long time, the graph start back at the beginning again.

peter

On Tue, May 12, 2009 at 6:24 PM, sebb seb...@gmail.com wrote:
 On 11/05/2009, drubix andrew.schr...@gmail.com wrote:

  Hi,

  When using the Graph Results listener, if the plot goes off the end of the X
  axis it continues to plot at time 0 again.  This means that there are 2 or
  more lines running over the top of each other, all of the same colour.  This
  produces some pretty nasty graphs which are difficult to read and are
  nonsensical.  I've attached an example:

  http://www.nabble.com/file/p23477206/graph.jpg

  I assume this isn't supposed to happen... does anybody know why mine is
  doing it?

 Not sure why this is happening; may be a threading issue.

 I suggest you save the samples to a JTL file - CSV mode is cheapest -
 and reload in the Listener after the test run. [This will be done in a
 single thread.]

 If it still causes wrap-round, then please raise a Bugzilla issue and
 attach the JTL file.

  Thanks,

  Drew

 --
  View this message in context: 
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
  Sent from the JMeter - User mailing list archive at Nabble.com.


  -
  To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
  For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



Re: Graph Results overlap

2009-05-12 Thread drubix

I've had this problem both when running the tests in GUI mode and non-GUI
mode (and opening the JTL file after the test has completed).  I'm away from
work today and don't have any of my testing files but I'll upload a bugzilla
report tomorrow if that is not the intended functionality.

Drew


Peter Lin wrote:
 
 Isn't that how it has always worked. If the test plan is large and
 runs for a long time, the graph start back at the beginning again.
 
 peter
 
 On Tue, May 12, 2009 at 6:24 PM, sebb seb...@gmail.com wrote:
 On 11/05/2009, drubix andrew.schr...@gmail.com wrote:

  Hi,

  When using the Graph Results listener, if the plot goes off the end of
 the X
  axis it continues to plot at time 0 again.  This means that there are 2
 or
  more lines running over the top of each other, all of the same colour.
  This
  produces some pretty nasty graphs which are difficult to read and are
  nonsensical.  I've attached an example:

  http://www.nabble.com/file/p23477206/graph.jpg

  I assume this isn't supposed to happen... does anybody know why mine is
  doing it?

 Not sure why this is happening; may be a threading issue.

 I suggest you save the samples to a JTL file - CSV mode is cheapest -
 and reload in the Listener after the test run. [This will be done in a
 single thread.]

 If it still causes wrap-round, then please raise a Bugzilla issue and
 attach the JTL file.

  Thanks,

  Drew

 --
  View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
  Sent from the JMeter - User mailing list archive at Nabble.com.


  -
  To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
  For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org


 
 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org
 
 
 

-- 
View this message in context: 
http://www.nabble.com/Graph-Results-overlap-tp23477206p23513354.html
Sent from the JMeter - User mailing list archive at Nabble.com.


-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



Re: Graph Results overlap

2009-05-12 Thread Peter Lin
I could be wrong, but it has always been that way. The default graph
listener has a fixed width, so once it reaches the end it just
continues at the beginning again.

peter

On Tue, May 12, 2009 at 8:29 PM, drubix andrew.schr...@gmail.com wrote:

 I've had this problem both when running the tests in GUI mode and non-GUI
 mode (and opening the JTL file after the test has completed).  I'm away from
 work today and don't have any of my testing files but I'll upload a bugzilla
 report tomorrow if that is not the intended functionality.

 Drew


 Peter Lin wrote:

 Isn't that how it has always worked. If the test plan is large and
 runs for a long time, the graph start back at the beginning again.

 peter

 On Tue, May 12, 2009 at 6:24 PM, sebb seb...@gmail.com wrote:
 On 11/05/2009, drubix andrew.schr...@gmail.com wrote:

  Hi,

  When using the Graph Results listener, if the plot goes off the end of
 the X
  axis it continues to plot at time 0 again.  This means that there are 2
 or
  more lines running over the top of each other, all of the same colour.
  This
  produces some pretty nasty graphs which are difficult to read and are
  nonsensical.  I've attached an example:

  http://www.nabble.com/file/p23477206/graph.jpg

  I assume this isn't supposed to happen... does anybody know why mine is
  doing it?

 Not sure why this is happening; may be a threading issue.

 I suggest you save the samples to a JTL file - CSV mode is cheapest -
 and reload in the Listener after the test run. [This will be done in a
 single thread.]

 If it still causes wrap-round, then please raise a Bugzilla issue and
 attach the JTL file.

  Thanks,

  Drew

 --
  View this message in context:
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23477206.html
  Sent from the JMeter - User mailing list archive at Nabble.com.


  -
  To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
  For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org




 --
 View this message in context: 
 http://www.nabble.com/Graph-Results-overlap-tp23477206p23513354.html
 Sent from the JMeter - User mailing list archive at Nabble.com.


 -
 To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
 For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org



-
To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org
For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org