You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by Will Meyer <wi...@bracketstyle.org> on 2005/11/23 17:08:58 UTC
Newbie Q re merging JTL files
I am aware of the remote testing capability, though for distributing test
clietns across the net it is not reasonable to have RMI connectivyt between
them. Rather, I am running multiple independent tests (all in console mode)
to generate maximum possible real-world load against a server. I was about
to wrtie some scritp to merge the JTL files together (CSV versions, since I
am doing stuff with the results in Excel), but it just ocurred to me that
this may already have been done (a tool like mergelog for clf files).
I realize that without true connectivity between the machines it won't be
possible for the tests to be truly accurate (due to time sync issues) but am
assuming that as long as I can increment the sampler numbers appropriately
(so there are not multiple instances of 1-N), I should have a pretty
meaningful aggregate result.
Does anyone have any pointers?
Thanks,
Will
Re: Newbie Q re merging JTL files
Posted by sebb <se...@gmail.com>.
Thanks - have fun! ...
If you want to post the code, please create a Buzilla enhancement
request, and then attach the code to that. Code sent to the list tends
to get lost/mangled/forgotten/etc.
S.
On 28/11/05, Will Meyer <wi...@bracketstyle.org> wrote:
> Oh ok, I thought there was a sample counter in there as well, but it looks
> like just the start time and sampled time, so it's as you say, pretty
> simple. I'll let the list know once I get around to doing this, I will
> probably do it in Java (since JMeter users obviously already have it) and
> not do anything too fancy re the algorithm.
>
> Will
>
>
> On 11/23/05, sebb <se...@gmail.com> wrote:
> >
> > On 23/11/05, Will Meyer <wi...@bracketstyle.org> wrote:
> > > That's a good idea, thanks. So has anyone already written the "merge N
> > JTL
> > > files together" script? I'd prefer to have them together anyway for
> > > analysis with other tools beyond the jmeter visualizers (though they're
> > > great).
> >
> > In the case of CSV files, all you need to do is concatenate and then sort
> > them.
> >
> > What language were you thinking of using for the script?
> >
> > > Thanks!
> > >
> > > Will
> > >
> > > On 11/23/05, Peter Lin <wo...@gmail.com> wrote:
> > > >
> > > >
> > > > one option is to use the timer functionality in the threadGroup to
> > start
> > > > the test. this does mean all the systems have to be sync-ed. One other
> > > > option might be to extend the listener to allow multiple file
> > selections.
> > > > for example, like extending ResultCollector so that any listener can
> > take
> > > > multiple .JTL files
> > > >
> > > > peter
> > > >
> > > >
> > > > On 11/23/05, Will Meyer <wi...@bracketstyle.org> wrote:
> > > > >
> > > > > I am aware of the remote testing capability, though for distributing
> > > > > test
> > > > > clietns across the net it is not reasonable to have RMI connectivyt
> > > > > between
> > > > > them. Rather, I am running multiple independent tests (all in
> > console
> > > > > mode)
> > > > > to generate maximum possible real-world load against a server. I
> > was
> > > > > about
> > > > > to wrtie some scritp to merge the JTL files together (CSV versions,
> > > > > since I
> > > > > am doing stuff with the results in Excel), but it just ocurred to me
> > > > > that
> > > > > this may already have been done (a tool like mergelog for clf
> > files).
> > > > >
> > > > > I realize that without true connectivity between the machines it
> > won't
> > > > > be
> > > > > possible for the tests to be truly accurate (due to time sync
> > issues)
> > > > > but am
> > > > > assuming that as long as I can increment the sampler numbers
> > > > > appropriately
> > > > > (so there are not multiple instances of 1-N), I should have a pretty
> >
> > > > > meaningful aggregate result.
> > > > >
> > > > > Does anyone have any pointers?
> > > > >
> > > > > Thanks,
> > > > >
> > > > > Will
> > > > >
> > > > >
> > > >
> > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
> >
> >
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Newbie Q re merging JTL files
Posted by Will Meyer <wi...@bracketstyle.org>.
Oh ok, I thought there was a sample counter in there as well, but it looks
like just the start time and sampled time, so it's as you say, pretty
simple. I'll let the list know once I get around to doing this, I will
probably do it in Java (since JMeter users obviously already have it) and
not do anything too fancy re the algorithm.
Will
On 11/23/05, sebb <se...@gmail.com> wrote:
>
> On 23/11/05, Will Meyer <wi...@bracketstyle.org> wrote:
> > That's a good idea, thanks. So has anyone already written the "merge N
> JTL
> > files together" script? I'd prefer to have them together anyway for
> > analysis with other tools beyond the jmeter visualizers (though they're
> > great).
>
> In the case of CSV files, all you need to do is concatenate and then sort
> them.
>
> What language were you thinking of using for the script?
>
> > Thanks!
> >
> > Will
> >
> > On 11/23/05, Peter Lin <wo...@gmail.com> wrote:
> > >
> > >
> > > one option is to use the timer functionality in the threadGroup to
> start
> > > the test. this does mean all the systems have to be sync-ed. One other
> > > option might be to extend the listener to allow multiple file
> selections.
> > > for example, like extending ResultCollector so that any listener can
> take
> > > multiple .JTL files
> > >
> > > peter
> > >
> > >
> > > On 11/23/05, Will Meyer <wi...@bracketstyle.org> wrote:
> > > >
> > > > I am aware of the remote testing capability, though for distributing
> > > > test
> > > > clietns across the net it is not reasonable to have RMI connectivyt
> > > > between
> > > > them. Rather, I am running multiple independent tests (all in
> console
> > > > mode)
> > > > to generate maximum possible real-world load against a server. I
> was
> > > > about
> > > > to wrtie some scritp to merge the JTL files together (CSV versions,
> > > > since I
> > > > am doing stuff with the results in Excel), but it just ocurred to me
> > > > that
> > > > this may already have been done (a tool like mergelog for clf
> files).
> > > >
> > > > I realize that without true connectivity between the machines it
> won't
> > > > be
> > > > possible for the tests to be truly accurate (due to time sync
> issues)
> > > > but am
> > > > assuming that as long as I can increment the sampler numbers
> > > > appropriately
> > > > (so there are not multiple instances of 1-N), I should have a pretty
>
> > > > meaningful aggregate result.
> > > >
> > > > Does anyone have any pointers?
> > > >
> > > > Thanks,
> > > >
> > > > Will
> > > >
> > > >
> > >
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>
>
Re: Newbie Q re merging JTL files
Posted by sebb <se...@gmail.com>.
On 23/11/05, Will Meyer <wi...@bracketstyle.org> wrote:
> That's a good idea, thanks. So has anyone already written the "merge N JTL
> files together" script? I'd prefer to have them together anyway for
> analysis with other tools beyond the jmeter visualizers (though they're
> great).
In the case of CSV files, all you need to do is concatenate and then sort them.
What language were you thinking of using for the script?
> Thanks!
>
> Will
>
> On 11/23/05, Peter Lin <wo...@gmail.com> wrote:
> >
> >
> > one option is to use the timer functionality in the threadGroup to start
> > the test. this does mean all the systems have to be sync-ed. One other
> > option might be to extend the listener to allow multiple file selections.
> > for example, like extending ResultCollector so that any listener can take
> > multiple .JTL files
> >
> > peter
> >
> >
> > On 11/23/05, Will Meyer <wi...@bracketstyle.org> wrote:
> > >
> > > I am aware of the remote testing capability, though for distributing
> > > test
> > > clietns across the net it is not reasonable to have RMI connectivyt
> > > between
> > > them. Rather, I am running multiple independent tests (all in console
> > > mode)
> > > to generate maximum possible real-world load against a server. I was
> > > about
> > > to wrtie some scritp to merge the JTL files together (CSV versions,
> > > since I
> > > am doing stuff with the results in Excel), but it just ocurred to me
> > > that
> > > this may already have been done (a tool like mergelog for clf files).
> > >
> > > I realize that without true connectivity between the machines it won't
> > > be
> > > possible for the tests to be truly accurate (due to time sync issues)
> > > but am
> > > assuming that as long as I can increment the sampler numbers
> > > appropriately
> > > (so there are not multiple instances of 1-N), I should have a pretty
> > > meaningful aggregate result.
> > >
> > > Does anyone have any pointers?
> > >
> > > Thanks,
> > >
> > > Will
> > >
> > >
> >
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
Re: Newbie Q re merging JTL files
Posted by Will Meyer <wi...@bracketstyle.org>.
That's a good idea, thanks. So has anyone already written the "merge N JTL
files together" script? I'd prefer to have them together anyway for
analysis with other tools beyond the jmeter visualizers (though they're
great).
Thanks!
Will
On 11/23/05, Peter Lin <wo...@gmail.com> wrote:
>
>
> one option is to use the timer functionality in the threadGroup to start
> the test. this does mean all the systems have to be sync-ed. One other
> option might be to extend the listener to allow multiple file selections.
> for example, like extending ResultCollector so that any listener can take
> multiple .JTL files
>
> peter
>
>
> On 11/23/05, Will Meyer <wi...@bracketstyle.org> wrote:
> >
> > I am aware of the remote testing capability, though for distributing
> > test
> > clietns across the net it is not reasonable to have RMI connectivyt
> > between
> > them. Rather, I am running multiple independent tests (all in console
> > mode)
> > to generate maximum possible real-world load against a server. I was
> > about
> > to wrtie some scritp to merge the JTL files together (CSV versions,
> > since I
> > am doing stuff with the results in Excel), but it just ocurred to me
> > that
> > this may already have been done (a tool like mergelog for clf files).
> >
> > I realize that without true connectivity between the machines it won't
> > be
> > possible for the tests to be truly accurate (due to time sync issues)
> > but am
> > assuming that as long as I can increment the sampler numbers
> > appropriately
> > (so there are not multiple instances of 1-N), I should have a pretty
> > meaningful aggregate result.
> >
> > Does anyone have any pointers?
> >
> > Thanks,
> >
> > Will
> >
> >
>
Re: Newbie Q re merging JTL files
Posted by Peter Lin <wo...@gmail.com>.
one option is to use the timer functionality in the threadGroup to start the
test. this does mean all the systems have to be sync-ed. One other option
might be to extend the listener to allow multiple file selections. for
example, like extending ResultCollector so that any listener can take
multiple .JTL files
peter
On 11/23/05, Will Meyer <wi...@bracketstyle.org> wrote:
>
> I am aware of the remote testing capability, though for distributing test
> clietns across the net it is not reasonable to have RMI connectivyt
> between
> them. Rather, I am running multiple independent tests (all in console
> mode)
> to generate maximum possible real-world load against a server. I was
> about
> to wrtie some scritp to merge the JTL files together (CSV versions, since
> I
> am doing stuff with the results in Excel), but it just ocurred to me that
> this may already have been done (a tool like mergelog for clf files).
>
> I realize that without true connectivity between the machines it won't be
> possible for the tests to be truly accurate (due to time sync issues) but
> am
> assuming that as long as I can increment the sampler numbers appropriately
> (so there are not multiple instances of 1-N), I should have a pretty
> meaningful aggregate result.
>
> Does anyone have any pointers?
>
> Thanks,
>
> Will
>
>