You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@jmeter.apache.org by sasidharsmit <sa...@gmail.com> on 2011/11/15 15:44:49 UTC
JMeter reporting higher response times
Dear All,
I'm using JMeter v2.5.1 r1177103 to automate one of my business flows, in a
windows 7 machine. The sampler I use is a custom plugin sampler "jp@gc -
HTTP Raw Request", that I downloaded from the below URL.
http://code.google.com/p/jmeter-plugins/
The business flow is simple.
a) I send a http raw request (requesting account details for a particular
user) and
b) I expect a response (with the details of the user).
I have completed automating the application and my script completes the flow
successfully. I get the response I expect and no errors are observed.
Everything is fine, except that JMeter is reporting response times higher
than what I observe with any other performance testing tool.
I have tried automating the same business flow via the following tools:
1. HTTP Requester
2. HP Performance Center
3. JMeter
The response times I get via the first two tools, are similar, around 40 ms,
on an average. But the response times observed via JMeter is just over 2000
ms (for every sample).
I'm running the test from my local desktop for a single user, in all 3
cases. The network capacity and the server capacity are the same in all 3
cases.
I have tried executing the tests with each of these 3 tools several times
and yet get the same kind of result. I doubt that there is something wrong
with the way I've designed my JMeter test plan.
I have attached the screenshots of JMeter and HTTP Requester results. As can
be seen in the screenshots,
a) I have executed 8 tests with HTTP Requester, each returning 518 bytes
of data, and takes around 40 ms
b) I have executed close to 100 tests with JMeter, each returning 518
bytes of data, and takes over 2000 ms, every time
One point that interests me is the difference between "Load time" and
"Latency" in the JMeter screenshots.
I could see that the latency is around 10-15 ms for each of the samples but
the load time is over 2000 ms in each sample. And, when I minus the latency
from the load time, almost every time I get a value of around 1998-2000 ms.
I do not know, exactly, what this means. But I doubt that the load time
includes some constant time which results in such high response times.
Can any of you help me find the problem area?
http://jmeter.512774.n5.nabble.com/file/n4994460/httprequestor_screenshot.png
httprequestor_screenshot.png
http://jmeter.512774.n5.nabble.com/file/n4994460/jmeter_screenshot.png
jmeter_screenshot.png
Thanks and Regards,
Sasidhar Sekar
--
View this message in context: http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp4994460p4994460.html
Sent from the JMeter - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by Mariusz W <ma...@gmail.com>.
Hi,
The meaning od "elapsedTime" in Requester should be checked imho. E.g. on
http://jmeter.512774.n5.nabble.com/file/n4994460/httprequestor_screenshot.pngthere
is almost the same time on 3:10:19 and 3:10:24 sample - and
keep-alive is 2 sec. so after first sample tcp connection should be closed
by server and second sample should be bigger then 40ms ( eg. google.com 3
way-handshake is about 40ms) so in your case should be about 80 in second
(3:10:24) sample. Maybe Requester don't count handshake time - only time
from first byte of response to last byte of response (==elapsedTime).
I think that you should verify time using Fiddler or Firebug yet.
Regards,
Mariusz
On Tue, Nov 15, 2011 at 3:44 PM, sasidharsmit <sa...@gmail.com>wrote:
> Dear All,
> I'm using JMeter v2.5.1 r1177103 to automate one of my business flows, in a
> windows 7 machine. The sampler I use is a custom plugin sampler "jp@gc -
> HTTP Raw Request", that I downloaded from the below URL.
>
> http://code.google.com/p/jmeter-plugins/
>
> The business flow is simple.
> a) I send a http raw request (requesting account details for a particular
> user) and
> b) I expect a response (with the details of the user).
>
> I have completed automating the application and my script completes the
> flow
> successfully. I get the response I expect and no errors are observed.
>
> Everything is fine, except that JMeter is reporting response times higher
> than what I observe with any other performance testing tool.
>
> I have tried automating the same business flow via the following tools:
>
> 1. HTTP Requester
> 2. HP Performance Center
> 3. JMeter
>
> The response times I get via the first two tools, are similar, around 40
> ms,
> on an average. But the response times observed via JMeter is just over 2000
> ms (for every sample).
>
> I'm running the test from my local desktop for a single user, in all 3
> cases. The network capacity and the server capacity are the same in all 3
> cases.
>
> I have tried executing the tests with each of these 3 tools several times
> and yet get the same kind of result. I doubt that there is something wrong
> with the way I've designed my JMeter test plan.
>
> I have attached the screenshots of JMeter and HTTP Requester results. As
> can
> be seen in the screenshots,
>
> a) I have executed 8 tests with HTTP Requester, each returning 518 bytes
> of data, and takes around 40 ms
> b) I have executed close to 100 tests with JMeter, each returning 518
> bytes of data, and takes over 2000 ms, every time
>
> One point that interests me is the difference between "Load time" and
> "Latency" in the JMeter screenshots.
> I could see that the latency is around 10-15 ms for each of the samples but
> the load time is over 2000 ms in each sample. And, when I minus the latency
> from the load time, almost every time I get a value of around 1998-2000 ms.
>
> I do not know, exactly, what this means. But I doubt that the load time
> includes some constant time which results in such high response times.
>
> Can any of you help me find the problem area?
>
>
> http://jmeter.512774.n5.nabble.com/file/n4994460/httprequestor_screenshot.png
> httprequestor_screenshot.png
> http://jmeter.512774.n5.nabble.com/file/n4994460/jmeter_screenshot.png
> jmeter_screenshot.png
>
> Thanks and Regards,
> Sasidhar Sekar
>
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp4994460p4994460.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
Re: JMeter reporting higher response times
Posted by sebb <se...@gmail.com>.
On 15 November 2011 14:44, sasidharsmit <sa...@gmail.com> wrote:
> Dear All,
> I'm using JMeter v2.5.1 r1177103 to automate one of my business flows, in a
> windows 7 machine. The sampler I use is a custom plugin sampler "jp@gc -
> HTTP Raw Request", that I downloaded from the below URL.
>
> http://code.google.com/p/jmeter-plugins/
This is not a standard JMeter plugin.
> The business flow is simple.
> a) I send a http raw request (requesting account details for a particular
> user) and
> b) I expect a response (with the details of the user).
>
> I have completed automating the application and my script completes the flow
> successfully. I get the response I expect and no errors are observed.
>
> Everything is fine, except that JMeter is reporting response times higher
> than what I observe with any other performance testing tool.
The response times are being generated by the 3rd party tool.
> I have tried automating the same business flow via the following tools:
>
> 1. HTTP Requester
> 2. HP Performance Center
> 3. JMeter
>
> The response times I get via the first two tools, are similar, around 40 ms,
> on an average. But the response times observed via JMeter is just over 2000
> ms (for every sample).
>
> I'm running the test from my local desktop for a single user, in all 3
> cases. The network capacity and the server capacity are the same in all 3
> cases.
>
> I have tried executing the tests with each of these 3 tools several times
> and yet get the same kind of result. I doubt that there is something wrong
> with the way I've designed my JMeter test plan.
>
> I have attached the screenshots of JMeter and HTTP Requester results. As can
> be seen in the screenshots,
>
> a) I have executed 8 tests with HTTP Requester, each returning 518 bytes
> of data, and takes around 40 ms
> b) I have executed close to 100 tests with JMeter, each returning 518
> bytes of data, and takes over 2000 ms, every time
>
> One point that interests me is the difference between "Load time" and
> "Latency" in the JMeter screenshots.
> I could see that the latency is around 10-15 ms for each of the samples but
> the load time is over 2000 ms in each sample. And, when I minus the latency
> from the load time, almost every time I get a value of around 1998-2000 ms.
>
> I do not know, exactly, what this means. But I doubt that the load time
> includes some constant time which results in such high response times.
>
> Can any of you help me find the problem area?
I suggest you ask the author of the 3rd party tool.
> http://jmeter.512774.n5.nabble.com/file/n4994460/httprequestor_screenshot.png
> httprequestor_screenshot.png
> http://jmeter.512774.n5.nabble.com/file/n4994460/jmeter_screenshot.png
> jmeter_screenshot.png
>
> Thanks and Regards,
> Sasidhar Sekar
>
>
> --
> View this message in context: http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp4994460p4994460.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by Philippe Mouawad <ph...@gmail.com>.
Hi,
Why not try your test by removing any plugins.
Also maybe you should try it with last trunk build of this day.
Regards
Philippe
On Tue, Nov 15, 2011 at 5:13 PM, Adrian Speteanu <as...@gmail.com>wrote:
> Hi guys,
>
> I actually got into a similar situation as the OP mentioned, soon after the
> release of 2.5.1. Eventually I tried to see if JMeter version has something
> to do with it, but I got similar response times. The problem was specific
> to how the client application and jmeter work. Too specific to be able to
> enter into details. The weird part is that the network packages seemed
> identical between the client and jmeter. So the problem might be the plugin
> (or how the listener works when using that plugin).
>
> Good luck,
> Adrian S
>
> On Tue, Nov 15, 2011 at 5:58 PM, Robin D. Wilson <rw...@gmail.com>
> wrote:
>
> > Not sure if this will matter, but have you tried testing this in JMeter
> > 2.4.x?
> >
> > The reason I ask is that I have a bunch of test scripts that I regularly
> > run
> > in 2.4 r961953, and I have seen a significant decrease in performance of
> > these same test scripts (both unmodified, and switching to various
> flavors
> > of the new HTTP Sampler) when I switched to 2.5 and 2.5.1.
> >
> > If you can back-port your script to 2.4, I'd be curious if it performs
> any
> > better.
> >
> > --
> > Robin D. Wilson
> > Sr. Director of Web Development
> > KingsIsle Entertainment, Inc.
> > VOICE: 512-777-1861
> > www.KingsIsle.com
> >
> >
> > -----Original Message-----
> > From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> > Sent: Tuesday, November 15, 2011 9:08 AM
> > To: jmeter-user@jakarta.apache.org
> > Subject: Re: JMeter reporting higher response times
> >
> > I disabled everything other than the actual sampler. Still, the response
> > time is over 2000 ms. PF attached the screenshot.
> >
> > http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.png
> > jmeter_only_sampler.png
> >
> > Regards,
> > Sasidhar Sekar
> >
> > --
> > View this message in context:
> >
> >
> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp
> > 4994460p4994555.html
> > Sent from the JMeter - User mailing list archive at Nabble.com.
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
>
--
Cordialement.
Philippe Mouawad.
Re: JMeter reporting higher response times
Posted by Adrian Speteanu <as...@gmail.com>.
Hi guys,
I actually got into a similar situation as the OP mentioned, soon after the
release of 2.5.1. Eventually I tried to see if JMeter version has something
to do with it, but I got similar response times. The problem was specific
to how the client application and jmeter work. Too specific to be able to
enter into details. The weird part is that the network packages seemed
identical between the client and jmeter. So the problem might be the plugin
(or how the listener works when using that plugin).
Good luck,
Adrian S
On Tue, Nov 15, 2011 at 5:58 PM, Robin D. Wilson <rw...@gmail.com> wrote:
> Not sure if this will matter, but have you tried testing this in JMeter
> 2.4.x?
>
> The reason I ask is that I have a bunch of test scripts that I regularly
> run
> in 2.4 r961953, and I have seen a significant decrease in performance of
> these same test scripts (both unmodified, and switching to various flavors
> of the new HTTP Sampler) when I switched to 2.5 and 2.5.1.
>
> If you can back-port your script to 2.4, I'd be curious if it performs any
> better.
>
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> Sent: Tuesday, November 15, 2011 9:08 AM
> To: jmeter-user@jakarta.apache.org
> Subject: Re: JMeter reporting higher response times
>
> I disabled everything other than the actual sampler. Still, the response
> time is over 2000 ms. PF attached the screenshot.
>
> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.png
> jmeter_only_sampler.png
>
> Regards,
> Sasidhar Sekar
>
> --
> View this message in context:
>
> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp
> 4994460p4994555.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
I've opened a bug: https://issues.apache.org/bugzilla/show_bug.cgi?id=52189
The file is attached there. I also attached both of the previous screen
images.
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: sebb [mailto:sebbaz@gmail.com]
Sent: Tuesday, November 15, 2011 2:12 PM
To: JMeter Users List
Subject: Re: JMeter reporting higher response times
On 15 November 2011 19:36, Robin D. Wilson <rw...@gmail.com> wrote:
> Both tests used HC 3.1, since HC 4 didn't work for the cookies (in the
> version I have of JM 2.5.1).
>
> Download Embedded Resources is disabled for the test. (So concurrent
> download is N/A.)
>
> Keepalive is enabled for all requests.
>
> I've placed my test in a zip file on Rapidshare:
> https://rapidshare.com/files/911894117/JMeterCompare.zip?bin=1
>
> (I've never used RS before, so hopefully that works for you.)
Does not seem to work for me.
An alternative is to create a Bugzilla issue and attach the file there.
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
> Sent: Tuesday, November 15, 2011 12:22 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> Can you provide your test plan or give more details about the options
> you use in your sampler:
>
> - So you said HC 3.1 (regarding HC4 the cookie issue you report is
> fixed)
> - Do you use embedded resources download ?
> - Do you use concurrent download option ? size of pool ?
> - keepalive ?
>
>
> Regards
> Philippe
>
> On Tue, Nov 15, 2011 at 6:37 PM, sebb <se...@gmail.com> wrote:
>
>> On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
>> > Hopefully this list can handle images...
>>
>> No, it cannot.
>>
>> Use a public hosting service and include the URL in the e-mail.
>>
>> > I have included a couple screen snaps of what I'm seeing. I ran a
>> > test
>> that
>> > requests (GET) the 'homepage' from our test machine. Then it POSTs
>> > a
>> login
>> > (username + password + special token) to the login form. It is a
>> > very simplified test.
>> >
>> > When I run it on JMeter 2.4, I get throughput of 128.1. When I run
>> > it on JMeter 2.5.1 I get throughput of 79.5.
>> >
>> > The request averages show similar differences:
>> >
>> > JMeter 2.4 - average times are 744ms
>> > JMeter 2.5.1 - average times are 901ms
>> >
>> > Re-running the tests over and over give very similar results
>> > (there's
>> some
>> > variability in the response times of the server, so I had to run
>> > this a bunch of times to be sure that I wasn't just seeing an
>> > anomaly. The
>> results
>> > are consistent - each test run shows JMeter 2.4 running
>> > significantly
>> faster
>> > than JMeter 2.5.1.
>>
>> Do you use HTTP or HTTPS?
>>
>> Do both GET and POST show elapsed time increases?
>>
>> > BTW, just now, when I tried to switch this test to use the HTTP4
>> > client,
>> the
>> > cookies stopped working correctly. So I couldn't test that sampler.
>> > The above numbers are just for the HTTP3.1 client. I have a cookie
>> > value
>> called
>> > 'stk', and using the "${COOKIE_stk}" variable in the POST fails
>> > about
>> 90% of
>> > the time when I'm using the HTTP4 client. It works fine with HTTP3.1.
>>
>> There have been some fixes to HC4 since 2.5.1.
>>
>> > --
>> > Robin D. Wilson
>> > Sr. Director of Web Development
>> > KingsIsle Entertainment, Inc.
>> > VOICE: 512-777-1861
>> > www.KingsIsle.com
>> >
>> >
>> > -----Original Message-----
>> > From: sebb [mailto:sebbaz@gmail.com]
>> > Sent: Tuesday, November 15, 2011 10:09 AM
>> > To: JMeter Users List
>> > Subject: Re: JMeter reporting higher response times
>> >
>> > On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
>> >> Not sure if this will matter, but have you tried testing this in
>> >> JMeter 2.4.x?
>> >>
>> >> The reason I ask is that I have a bunch of test scripts that I
>> >> regularly run in 2.4 r961953, and I have seen a significant
>> >> decrease in performance of these same test scripts (both
>> >> unmodified, and switching to various flavors of the new HTTP
>> >> Sampler) when I switched to
>> > 2.5 and 2.5.1.
>> >
>> > Can you provide details of these issues please?
>> >
>> >> --
>> >> Robin D. Wilson
>> >> Sr. Director of Web Development
>> >> KingsIsle Entertainment, Inc.
>> >> VOICE: 512-777-1861
>> >> www.KingsIsle.com
>> >>
>> >>
>> >> -----Original Message-----
>> >> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
>> >> Sent: Tuesday, November 15, 2011 9:08 AM
>> >> To: jmeter-user@jakarta.apache.org
>> >> Subject: Re: JMeter reporting higher response times
>> >>
>> >> I disabled everything other than the actual sampler. Still, the
>> >> response time is over 2000 ms. PF attached the screenshot.
>> >>
>> >> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampl
>> >> e
>> >> r.p
>> >> ng
>> >> jmeter_only_sampler.png
>> >>
>> >> Regards,
>> >> Sasidhar Sekar
>> >>
>> >> --
>> >> View this message in context:
>> >> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-respons
>> >> e
>> >> -ti
>> >> mes-tp
>> >> 4994460p4994555.html
>> >> Sent from the JMeter - User mailing list archive at Nabble.com.
>> >>
>> >> ------------------------------------------------------------------
>> >> -
>> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>
>> >>
>> >> ------------------------------------------------------------------
>> >> -
>> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>
>> >>
>> >
>> > -------------------------------------------------------------------
>> > -
>> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> > For additional commands, e-mail: user-help@jmeter.apache.org
>> >
>> >
>> >
>> > -------------------------------------------------------------------
>> > -
>> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> > For additional commands, e-mail: user-help@jmeter.apache.org
>> >
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>
>
> --
> Cordialement.
> Philippe Mouawad.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by Philippe Mouawad <ph...@gmail.com>.
It worked for me, but we cannot use it as web site is not public
On Tue, Nov 15, 2011 at 9:11 PM, sebb <se...@gmail.com> wrote:
> On 15 November 2011 19:36, Robin D. Wilson <rw...@gmail.com> wrote:
> > Both tests used HC 3.1, since HC 4 didn't work for the cookies (in the
> > version I have of JM 2.5.1).
> >
> > Download Embedded Resources is disabled for the test. (So concurrent
> > download is N/A.)
> >
> > Keepalive is enabled for all requests.
> >
> > I've placed my test in a zip file on Rapidshare:
> > https://rapidshare.com/files/911894117/JMeterCompare.zip?bin=1
> >
> > (I've never used RS before, so hopefully that works for you.)
>
> Does not seem to work for me.
>
> An alternative is to create a Bugzilla issue and attach the file there.
>
> > --
> > Robin D. Wilson
> > Sr. Director of Web Development
> > KingsIsle Entertainment, Inc.
> > VOICE: 512-777-1861
> > www.KingsIsle.com
> >
> >
> > -----Original Message-----
> > From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
> > Sent: Tuesday, November 15, 2011 12:22 PM
> > To: JMeter Users List
> > Subject: Re: JMeter reporting higher response times
> >
> > Can you provide your test plan or give more details about the options you
> > use in your sampler:
> >
> > - So you said HC 3.1 (regarding HC4 the cookie issue you report is
> fixed)
> > - Do you use embedded resources download ?
> > - Do you use concurrent download option ? size of pool ?
> > - keepalive ?
> >
> >
> > Regards
> > Philippe
> >
> > On Tue, Nov 15, 2011 at 6:37 PM, sebb <se...@gmail.com> wrote:
> >
> >> On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
> >> > Hopefully this list can handle images...
> >>
> >> No, it cannot.
> >>
> >> Use a public hosting service and include the URL in the e-mail.
> >>
> >> > I have included a couple screen snaps of what I'm seeing. I ran a
> >> > test
> >> that
> >> > requests (GET) the 'homepage' from our test machine. Then it POSTs a
> >> login
> >> > (username + password + special token) to the login form. It is a
> >> > very simplified test.
> >> >
> >> > When I run it on JMeter 2.4, I get throughput of 128.1. When I run
> >> > it on JMeter 2.5.1 I get throughput of 79.5.
> >> >
> >> > The request averages show similar differences:
> >> >
> >> > JMeter 2.4 - average times are 744ms
> >> > JMeter 2.5.1 - average times are 901ms
> >> >
> >> > Re-running the tests over and over give very similar results
> >> > (there's
> >> some
> >> > variability in the response times of the server, so I had to run
> >> > this a bunch of times to be sure that I wasn't just seeing an
> >> > anomaly. The
> >> results
> >> > are consistent - each test run shows JMeter 2.4 running
> >> > significantly
> >> faster
> >> > than JMeter 2.5.1.
> >>
> >> Do you use HTTP or HTTPS?
> >>
> >> Do both GET and POST show elapsed time increases?
> >>
> >> > BTW, just now, when I tried to switch this test to use the HTTP4
> >> > client,
> >> the
> >> > cookies stopped working correctly. So I couldn't test that sampler.
> >> > The above numbers are just for the HTTP3.1 client. I have a cookie
> >> > value
> >> called
> >> > 'stk', and using the "${COOKIE_stk}" variable in the POST fails
> >> > about
> >> 90% of
> >> > the time when I'm using the HTTP4 client. It works fine with HTTP3.1.
> >>
> >> There have been some fixes to HC4 since 2.5.1.
> >>
> >> > --
> >> > Robin D. Wilson
> >> > Sr. Director of Web Development
> >> > KingsIsle Entertainment, Inc.
> >> > VOICE: 512-777-1861
> >> > www.KingsIsle.com
> >> >
> >> >
> >> > -----Original Message-----
> >> > From: sebb [mailto:sebbaz@gmail.com]
> >> > Sent: Tuesday, November 15, 2011 10:09 AM
> >> > To: JMeter Users List
> >> > Subject: Re: JMeter reporting higher response times
> >> >
> >> > On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com>
> wrote:
> >> >> Not sure if this will matter, but have you tried testing this in
> >> >> JMeter 2.4.x?
> >> >>
> >> >> The reason I ask is that I have a bunch of test scripts that I
> >> >> regularly run in 2.4 r961953, and I have seen a significant
> >> >> decrease in performance of these same test scripts (both
> >> >> unmodified, and switching to various flavors of the new HTTP
> >> >> Sampler) when I switched to
> >> > 2.5 and 2.5.1.
> >> >
> >> > Can you provide details of these issues please?
> >> >
> >> >> --
> >> >> Robin D. Wilson
> >> >> Sr. Director of Web Development
> >> >> KingsIsle Entertainment, Inc.
> >> >> VOICE: 512-777-1861
> >> >> www.KingsIsle.com
> >> >>
> >> >>
> >> >> -----Original Message-----
> >> >> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> >> >> Sent: Tuesday, November 15, 2011 9:08 AM
> >> >> To: jmeter-user@jakarta.apache.org
> >> >> Subject: Re: JMeter reporting higher response times
> >> >>
> >> >> I disabled everything other than the actual sampler. Still, the
> >> >> response time is over 2000 ms. PF attached the screenshot.
> >> >>
> >> >> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sample
> >> >> r.p
> >> >> ng
> >> >> jmeter_only_sampler.png
> >> >>
> >> >> Regards,
> >> >> Sasidhar Sekar
> >> >>
> >> >> --
> >> >> View this message in context:
> >> >> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response
> >> >> -ti
> >> >> mes-tp
> >> >> 4994460p4994555.html
> >> >> Sent from the JMeter - User mailing list archive at Nabble.com.
> >> >>
> >> >> -------------------------------------------------------------------
> >> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>
> >> >>
> >> >> -------------------------------------------------------------------
> >> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>
> >> >>
> >> >
> >> > --------------------------------------------------------------------
> >> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> > For additional commands, e-mail: user-help@jmeter.apache.org
> >> >
> >> >
> >> >
> >> > --------------------------------------------------------------------
> >> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> > For additional commands, e-mail: user-help@jmeter.apache.org
> >> >
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >
> >
> > --
> > Cordialement.
> > Philippe Mouawad.
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
--
Cordialement.
Philippe Mouawad.
Re: JMeter reporting higher response times
Posted by sebb <se...@gmail.com>.
On 15 November 2011 19:36, Robin D. Wilson <rw...@gmail.com> wrote:
> Both tests used HC 3.1, since HC 4 didn't work for the cookies (in the
> version I have of JM 2.5.1).
>
> Download Embedded Resources is disabled for the test. (So concurrent
> download is N/A.)
>
> Keepalive is enabled for all requests.
>
> I've placed my test in a zip file on Rapidshare:
> https://rapidshare.com/files/911894117/JMeterCompare.zip?bin=1
>
> (I've never used RS before, so hopefully that works for you.)
Does not seem to work for me.
An alternative is to create a Bugzilla issue and attach the file there.
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
> Sent: Tuesday, November 15, 2011 12:22 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> Can you provide your test plan or give more details about the options you
> use in your sampler:
>
> - So you said HC 3.1 (regarding HC4 the cookie issue you report is fixed)
> - Do you use embedded resources download ?
> - Do you use concurrent download option ? size of pool ?
> - keepalive ?
>
>
> Regards
> Philippe
>
> On Tue, Nov 15, 2011 at 6:37 PM, sebb <se...@gmail.com> wrote:
>
>> On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
>> > Hopefully this list can handle images...
>>
>> No, it cannot.
>>
>> Use a public hosting service and include the URL in the e-mail.
>>
>> > I have included a couple screen snaps of what I'm seeing. I ran a
>> > test
>> that
>> > requests (GET) the 'homepage' from our test machine. Then it POSTs a
>> login
>> > (username + password + special token) to the login form. It is a
>> > very simplified test.
>> >
>> > When I run it on JMeter 2.4, I get throughput of 128.1. When I run
>> > it on JMeter 2.5.1 I get throughput of 79.5.
>> >
>> > The request averages show similar differences:
>> >
>> > JMeter 2.4 - average times are 744ms
>> > JMeter 2.5.1 - average times are 901ms
>> >
>> > Re-running the tests over and over give very similar results
>> > (there's
>> some
>> > variability in the response times of the server, so I had to run
>> > this a bunch of times to be sure that I wasn't just seeing an
>> > anomaly. The
>> results
>> > are consistent - each test run shows JMeter 2.4 running
>> > significantly
>> faster
>> > than JMeter 2.5.1.
>>
>> Do you use HTTP or HTTPS?
>>
>> Do both GET and POST show elapsed time increases?
>>
>> > BTW, just now, when I tried to switch this test to use the HTTP4
>> > client,
>> the
>> > cookies stopped working correctly. So I couldn't test that sampler.
>> > The above numbers are just for the HTTP3.1 client. I have a cookie
>> > value
>> called
>> > 'stk', and using the "${COOKIE_stk}" variable in the POST fails
>> > about
>> 90% of
>> > the time when I'm using the HTTP4 client. It works fine with HTTP3.1.
>>
>> There have been some fixes to HC4 since 2.5.1.
>>
>> > --
>> > Robin D. Wilson
>> > Sr. Director of Web Development
>> > KingsIsle Entertainment, Inc.
>> > VOICE: 512-777-1861
>> > www.KingsIsle.com
>> >
>> >
>> > -----Original Message-----
>> > From: sebb [mailto:sebbaz@gmail.com]
>> > Sent: Tuesday, November 15, 2011 10:09 AM
>> > To: JMeter Users List
>> > Subject: Re: JMeter reporting higher response times
>> >
>> > On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
>> >> Not sure if this will matter, but have you tried testing this in
>> >> JMeter 2.4.x?
>> >>
>> >> The reason I ask is that I have a bunch of test scripts that I
>> >> regularly run in 2.4 r961953, and I have seen a significant
>> >> decrease in performance of these same test scripts (both
>> >> unmodified, and switching to various flavors of the new HTTP
>> >> Sampler) when I switched to
>> > 2.5 and 2.5.1.
>> >
>> > Can you provide details of these issues please?
>> >
>> >> --
>> >> Robin D. Wilson
>> >> Sr. Director of Web Development
>> >> KingsIsle Entertainment, Inc.
>> >> VOICE: 512-777-1861
>> >> www.KingsIsle.com
>> >>
>> >>
>> >> -----Original Message-----
>> >> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
>> >> Sent: Tuesday, November 15, 2011 9:08 AM
>> >> To: jmeter-user@jakarta.apache.org
>> >> Subject: Re: JMeter reporting higher response times
>> >>
>> >> I disabled everything other than the actual sampler. Still, the
>> >> response time is over 2000 ms. PF attached the screenshot.
>> >>
>> >> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sample
>> >> r.p
>> >> ng
>> >> jmeter_only_sampler.png
>> >>
>> >> Regards,
>> >> Sasidhar Sekar
>> >>
>> >> --
>> >> View this message in context:
>> >> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response
>> >> -ti
>> >> mes-tp
>> >> 4994460p4994555.html
>> >> Sent from the JMeter - User mailing list archive at Nabble.com.
>> >>
>> >> -------------------------------------------------------------------
>> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>
>> >>
>> >> -------------------------------------------------------------------
>> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>
>> >>
>> >
>> > --------------------------------------------------------------------
>> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> > For additional commands, e-mail: user-help@jmeter.apache.org
>> >
>> >
>> >
>> > --------------------------------------------------------------------
>> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> > For additional commands, e-mail: user-help@jmeter.apache.org
>> >
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>
>
> --
> Cordialement.
> Philippe Mouawad.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
Both tests used HC 3.1, since HC 4 didn't work for the cookies (in the
version I have of JM 2.5.1).
Download Embedded Resources is disabled for the test. (So concurrent
download is N/A.)
Keepalive is enabled for all requests.
I've placed my test in a zip file on Rapidshare:
https://rapidshare.com/files/911894117/JMeterCompare.zip?bin=1
(I've never used RS before, so hopefully that works for you.)
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
Sent: Tuesday, November 15, 2011 12:22 PM
To: JMeter Users List
Subject: Re: JMeter reporting higher response times
Can you provide your test plan or give more details about the options you
use in your sampler:
- So you said HC 3.1 (regarding HC4 the cookie issue you report is fixed)
- Do you use embedded resources download ?
- Do you use concurrent download option ? size of pool ?
- keepalive ?
Regards
Philippe
On Tue, Nov 15, 2011 at 6:37 PM, sebb <se...@gmail.com> wrote:
> On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
> > Hopefully this list can handle images...
>
> No, it cannot.
>
> Use a public hosting service and include the URL in the e-mail.
>
> > I have included a couple screen snaps of what I'm seeing. I ran a
> > test
> that
> > requests (GET) the 'homepage' from our test machine. Then it POSTs a
> login
> > (username + password + special token) to the login form. It is a
> > very simplified test.
> >
> > When I run it on JMeter 2.4, I get throughput of 128.1. When I run
> > it on JMeter 2.5.1 I get throughput of 79.5.
> >
> > The request averages show similar differences:
> >
> > JMeter 2.4 - average times are 744ms
> > JMeter 2.5.1 - average times are 901ms
> >
> > Re-running the tests over and over give very similar results
> > (there's
> some
> > variability in the response times of the server, so I had to run
> > this a bunch of times to be sure that I wasn't just seeing an
> > anomaly. The
> results
> > are consistent - each test run shows JMeter 2.4 running
> > significantly
> faster
> > than JMeter 2.5.1.
>
> Do you use HTTP or HTTPS?
>
> Do both GET and POST show elapsed time increases?
>
> > BTW, just now, when I tried to switch this test to use the HTTP4
> > client,
> the
> > cookies stopped working correctly. So I couldn't test that sampler.
> > The above numbers are just for the HTTP3.1 client. I have a cookie
> > value
> called
> > 'stk', and using the "${COOKIE_stk}" variable in the POST fails
> > about
> 90% of
> > the time when I'm using the HTTP4 client. It works fine with HTTP3.1.
>
> There have been some fixes to HC4 since 2.5.1.
>
> > --
> > Robin D. Wilson
> > Sr. Director of Web Development
> > KingsIsle Entertainment, Inc.
> > VOICE: 512-777-1861
> > www.KingsIsle.com
> >
> >
> > -----Original Message-----
> > From: sebb [mailto:sebbaz@gmail.com]
> > Sent: Tuesday, November 15, 2011 10:09 AM
> > To: JMeter Users List
> > Subject: Re: JMeter reporting higher response times
> >
> > On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
> >> Not sure if this will matter, but have you tried testing this in
> >> JMeter 2.4.x?
> >>
> >> The reason I ask is that I have a bunch of test scripts that I
> >> regularly run in 2.4 r961953, and I have seen a significant
> >> decrease in performance of these same test scripts (both
> >> unmodified, and switching to various flavors of the new HTTP
> >> Sampler) when I switched to
> > 2.5 and 2.5.1.
> >
> > Can you provide details of these issues please?
> >
> >> --
> >> Robin D. Wilson
> >> Sr. Director of Web Development
> >> KingsIsle Entertainment, Inc.
> >> VOICE: 512-777-1861
> >> www.KingsIsle.com
> >>
> >>
> >> -----Original Message-----
> >> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> >> Sent: Tuesday, November 15, 2011 9:08 AM
> >> To: jmeter-user@jakarta.apache.org
> >> Subject: Re: JMeter reporting higher response times
> >>
> >> I disabled everything other than the actual sampler. Still, the
> >> response time is over 2000 ms. PF attached the screenshot.
> >>
> >> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sample
> >> r.p
> >> ng
> >> jmeter_only_sampler.png
> >>
> >> Regards,
> >> Sasidhar Sekar
> >>
> >> --
> >> View this message in context:
> >> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response
> >> -ti
> >> mes-tp
> >> 4994460p4994555.html
> >> Sent from the JMeter - User mailing list archive at Nabble.com.
> >>
> >> -------------------------------------------------------------------
> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >> -------------------------------------------------------------------
> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >
> > --------------------------------------------------------------------
> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
> >
> > --------------------------------------------------------------------
> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
--
Cordialement.
Philippe Mouawad.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by Philippe Mouawad <ph...@gmail.com>.
Can you provide your test plan or give more details about the options you
use in your sampler:
- So you said HC 3.1 (regarding HC4 the cookie issue you report is fixed)
- Do you use embedded resources download ?
- Do you use concurrent download option ? size of pool ?
- keepalive ?
Regards
Philippe
On Tue, Nov 15, 2011 at 6:37 PM, sebb <se...@gmail.com> wrote:
> On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
> > Hopefully this list can handle images...
>
> No, it cannot.
>
> Use a public hosting service and include the URL in the e-mail.
>
> > I have included a couple screen snaps of what I'm seeing. I ran a test
> that
> > requests (GET) the 'homepage' from our test machine. Then it POSTs a
> login
> > (username + password + special token) to the login form. It is a very
> > simplified test.
> >
> > When I run it on JMeter 2.4, I get throughput of 128.1. When I run it on
> > JMeter 2.5.1 I get throughput of 79.5.
> >
> > The request averages show similar differences:
> >
> > JMeter 2.4 - average times are 744ms
> > JMeter 2.5.1 - average times are 901ms
> >
> > Re-running the tests over and over give very similar results (there's
> some
> > variability in the response times of the server, so I had to run this a
> > bunch of times to be sure that I wasn't just seeing an anomaly. The
> results
> > are consistent - each test run shows JMeter 2.4 running significantly
> faster
> > than JMeter 2.5.1.
>
> Do you use HTTP or HTTPS?
>
> Do both GET and POST show elapsed time increases?
>
> > BTW, just now, when I tried to switch this test to use the HTTP4 client,
> the
> > cookies stopped working correctly. So I couldn't test that sampler. The
> > above numbers are just for the HTTP3.1 client. I have a cookie value
> called
> > 'stk', and using the "${COOKIE_stk}" variable in the POST fails about
> 90% of
> > the time when I'm using the HTTP4 client. It works fine with HTTP3.1.
>
> There have been some fixes to HC4 since 2.5.1.
>
> > --
> > Robin D. Wilson
> > Sr. Director of Web Development
> > KingsIsle Entertainment, Inc.
> > VOICE: 512-777-1861
> > www.KingsIsle.com
> >
> >
> > -----Original Message-----
> > From: sebb [mailto:sebbaz@gmail.com]
> > Sent: Tuesday, November 15, 2011 10:09 AM
> > To: JMeter Users List
> > Subject: Re: JMeter reporting higher response times
> >
> > On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
> >> Not sure if this will matter, but have you tried testing this in
> >> JMeter 2.4.x?
> >>
> >> The reason I ask is that I have a bunch of test scripts that I
> >> regularly run in 2.4 r961953, and I have seen a significant decrease
> >> in performance of these same test scripts (both unmodified, and
> >> switching to various flavors of the new HTTP Sampler) when I switched to
> > 2.5 and 2.5.1.
> >
> > Can you provide details of these issues please?
> >
> >> --
> >> Robin D. Wilson
> >> Sr. Director of Web Development
> >> KingsIsle Entertainment, Inc.
> >> VOICE: 512-777-1861
> >> www.KingsIsle.com
> >>
> >>
> >> -----Original Message-----
> >> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> >> Sent: Tuesday, November 15, 2011 9:08 AM
> >> To: jmeter-user@jakarta.apache.org
> >> Subject: Re: JMeter reporting higher response times
> >>
> >> I disabled everything other than the actual sampler. Still, the
> >> response time is over 2000 ms. PF attached the screenshot.
> >>
> >> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.p
> >> ng
> >> jmeter_only_sampler.png
> >>
> >> Regards,
> >> Sasidhar Sekar
> >>
> >> --
> >> View this message in context:
> >> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-ti
> >> mes-tp
> >> 4994460p4994555.html
> >> Sent from the JMeter - User mailing list archive at Nabble.com.
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
--
Cordialement.
Philippe Mouawad.
Re: JMeter reporting higher response times
Posted by sasidharsmit <sa...@gmail.com>.
Got my issue resolved.
For anyone using the jmeter plugins from the below URL:
http://code.google.com/p/jmeter-plugins/
"jp@gc - HTTP Raw Request" sampler does not support "Connection:
Keep-Alive". So, changing the header to "Connection: close" resolved my
issue.
Disabling Keep-Alive has its own implications. So, please understand the
implications before making this change.
Cheers,
Sasidhar Sekar
--
View this message in context: http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp4994460p4997473.html
Sent from the JMeter - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
OK, NOW I get it!
There was another ".zip" archive I needed (the 'lib')... You guys kept using
short-hand for stuff, and I was just too thick to grasp what you meant. I
thought when you were referring to "lib", you meant the "lib" folder in the
'apache-jmeter-r1202266_bin.zip' file. Not that there was another file:
'apache-jmeter-r1202266_lib.zip' that I needed...
Sorry about being so dense that I couldn't follow you.
Here's what I found when I ran the nightly build:
1) The cookies work again (thanks).
2) HTTP4 client is even slower than the HTTP3.1 client was (for both the GET
and the POST). I've uploaded another image to the bugzilla ticket.
https://issues.apache.org/bugzilla/show_bug.cgi?id=52189
3) Just for grins, I tried the "Java" implementation too... It performs
about the same (within the margin of error) as the HttpClient3.1
implementation under JM 2.5.1.
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
Sent: Tuesday, November 15, 2011 4:12 PM
To: JMeter Users List
Subject: Re: JMeter reporting higher response times
The bin Zip contains 2 jars but the lib zip contains dépendency jars so
merging both should do it.
You shouldn't use 2.5.1 jars because there has been upgrade ( jexl) If you
don't succeed making it work i Will try to provide you tomorrow a dist on a
download site.
Regards
Philippe
On Tuesday, November 15, 2011, Robin D. Wilson <rw...@gmail.com> wrote:
> That doesn't work either... When I do that, the 'lib' folder is
> missing a bunch of ".jar" files. The "lib" from the 'zip' file only
> contains 2
".jar"
> files, 'bshclient.jar' and 'jorphan.jar'.
>
> I tried it, but that just throws a stack trace (I would show you, but
> I'm pretty certain that it is related to the missing "jar" files). So
> I added back the 'difference' (e.g., the "jar" files that exist in the
> original JMeter2.5.1, but were missing from the JMeter2.5.1Dev after I
> wiped them all, but skipping the ones that were already there from the
> nightly build ".zip" file). The end result was the same error as before.
>
> Am I not getting the right nightly build file?
>
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
> Sent: Tuesday, November 15, 2011 3:42 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> In your procedure, before copying jars from new zip to 2.5.1 DEV,
> remove
all
> jars from your 2.5.1 DEV. (in lib, lib/ext and bin folder).
>
>
> On Tue, Nov 15, 2011 at 10:37 PM, Robin D. Wilson <rwilson2@gmail.com
>wrote:
>
>> I must be terribly thick headed...
>>
>> Here's what I did:
>>
>> Started with a folder called:
>>
>> JMeter2.5.1
>>
>> This contains the full (and functional) installation of JMeter 2.5.1.
>> When I launch the 'bin/jmeter.bat' file from this folder, it works
>> fine. The only modification to this folder is the
>> 'bin/user.properties'.
>>
>> I copied the JMeter2.5.1 folder to JMeter2.5.1Dev. I ran the
>> jmeter.bat file from this folder - it works fine.
>>
>> Then I extracted the 'apache-jmeter-r1202266_bin.zip' file, and
>> copied the ".jar" files from the "lib" folder to the 'JMeter2.5.1Dev/lib'
folder.
>> I copied the "lib/etx/*.jar" files to the 'JMeter2.5.1Dev/lib/ext'
folder.
>> I copied the "lib/junit/test.jar" file to the 'JMeter2.5.1Dev/lib/junit'
>> folder.
>> I copied the "bin/*" files (not the "bin/examples" folder) to the
>> 'JMeter2.5.1Dev/bin' folder.
>>
>> When I launch the 'bin/jmeter.bat' file, I get the following error:
>>
>> An error occurred: org/apache/commons/jexl2/JexlContext
>> errorlevel=1
>> Press any key to continue . . .
>>
>> (Of course, when I press a key, the cmd.exe window goes away.)
>>
>> I have only the JRE installed - it is possible that I need the full
>> JDK in order to get this working?
>>
>> --
>> Robin D. Wilson
>> Sr. Director of Web Development
>> KingsIsle Entertainment, Inc.
>> VOICE: 512-777-1861
>> www.KingsIsle.com
>>
>>
>> -----Original Message-----
>> From: sebb [mailto:sebbaz@gmail.com]
>> Sent: Tuesday, November 15, 2011 2:34 PM
>> To: JMeter Users List
>> Subject: Re: JMeter reporting higher response times
>>
>> On 15 November 2011 20:28, Robin D. Wilson <rw...@gmail.com> wrote:
>> > I must not be understanding what you are asking me to do...
>> >
>> > I downloaded the 'apache-jmeter-r1202266_bin.zip' file from the URL
>> below.
>> >
>> > Then I 'copied' my folder for the "JMeter2.5.1" install. (Then I
>> > verified that I can start JMeter from that new folder - it works
>> > fine.)
>> >
>> > Then I merged the "lib" folder from the above
>> > 'apache-jmeter-r1202266_bin.zip' file into the "lib" folder from
>> > the JMeter2.5.1_copy folder that I created.
>> >
>> > Now the JMeter in that folder won't start.
>> >
>> > I also tried merging all the files from ".zip" folder into the
>> > copy, no dice. What am I doing wrong?
>>
>> You need to unpack the _bin and _lib archives into the same directory
>> structure.
>>
>> There should be only a couple of duplicate text files at the top level.
>>
>> > --
>> > Robin D. Wilson
>> > Sr. Director of Web Development
>> > KingsIsle Entertainment, Inc.
>> > VOICE: 512-777-1861
>> > www.KingsIsle.com
>> >
>> >
>> > -----Original Message-----
>> > From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
>> > Sent: Tuesday, November 15, 2011 1:56 PM
>> > To: JMeter Users List
>> > Subject: Re: JMeter reporting higher response times
>> >
>> > You can get it here:
>> > https://builds.apache.org/job/JMeter-trunk/lastSuccessfulBuild/arti
>> > f
>> > ac
>> > t/trun
>> > k/dist/
>> >
>> > Merge lib zip into bin one .
>> >
>> >
--
Cordialement.
Philippe Mouawad.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by Philippe Mouawad <ph...@gmail.com>.
The bin Zip contains 2 jars but the lib zip contains dépendency jars so
merging both should do it.
You shouldn't use 2.5.1 jars because there has been upgrade ( jexl)
If you don't succeed making it work i Will try to provide you tomorrow a
dist on a download site.
Regards
Philippe
On Tuesday, November 15, 2011, Robin D. Wilson <rw...@gmail.com> wrote:
> That doesn't work either... When I do that, the 'lib' folder is missing a
> bunch of ".jar" files. The "lib" from the 'zip' file only contains 2
".jar"
> files, 'bshclient.jar' and 'jorphan.jar'.
>
> I tried it, but that just throws a stack trace (I would show you, but I'm
> pretty certain that it is related to the missing "jar" files). So I added
> back the 'difference' (e.g., the "jar" files that exist in the original
> JMeter2.5.1, but were missing from the JMeter2.5.1Dev after I wiped them
> all, but skipping the ones that were already there from the nightly build
> ".zip" file). The end result was the same error as before.
>
> Am I not getting the right nightly build file?
>
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
> Sent: Tuesday, November 15, 2011 3:42 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> In your procedure, before copying jars from new zip to 2.5.1 DEV, remove
all
> jars from your 2.5.1 DEV. (in lib, lib/ext and bin folder).
>
>
> On Tue, Nov 15, 2011 at 10:37 PM, Robin D. Wilson <rwilson2@gmail.com
>wrote:
>
>> I must be terribly thick headed...
>>
>> Here's what I did:
>>
>> Started with a folder called:
>>
>> JMeter2.5.1
>>
>> This contains the full (and functional) installation of JMeter 2.5.1.
>> When I launch the 'bin/jmeter.bat' file from this folder, it works
>> fine. The only modification to this folder is the
>> 'bin/user.properties'.
>>
>> I copied the JMeter2.5.1 folder to JMeter2.5.1Dev. I ran the
>> jmeter.bat file from this folder - it works fine.
>>
>> Then I extracted the 'apache-jmeter-r1202266_bin.zip' file, and copied
>> the ".jar" files from the "lib" folder to the 'JMeter2.5.1Dev/lib'
folder.
>> I copied the "lib/etx/*.jar" files to the 'JMeter2.5.1Dev/lib/ext'
folder.
>> I copied the "lib/junit/test.jar" file to the 'JMeter2.5.1Dev/lib/junit'
>> folder.
>> I copied the "bin/*" files (not the "bin/examples" folder) to the
>> 'JMeter2.5.1Dev/bin' folder.
>>
>> When I launch the 'bin/jmeter.bat' file, I get the following error:
>>
>> An error occurred: org/apache/commons/jexl2/JexlContext
>> errorlevel=1
>> Press any key to continue . . .
>>
>> (Of course, when I press a key, the cmd.exe window goes away.)
>>
>> I have only the JRE installed - it is possible that I need the full
>> JDK in order to get this working?
>>
>> --
>> Robin D. Wilson
>> Sr. Director of Web Development
>> KingsIsle Entertainment, Inc.
>> VOICE: 512-777-1861
>> www.KingsIsle.com
>>
>>
>> -----Original Message-----
>> From: sebb [mailto:sebbaz@gmail.com]
>> Sent: Tuesday, November 15, 2011 2:34 PM
>> To: JMeter Users List
>> Subject: Re: JMeter reporting higher response times
>>
>> On 15 November 2011 20:28, Robin D. Wilson <rw...@gmail.com> wrote:
>> > I must not be understanding what you are asking me to do...
>> >
>> > I downloaded the 'apache-jmeter-r1202266_bin.zip' file from the URL
>> below.
>> >
>> > Then I 'copied' my folder for the "JMeter2.5.1" install. (Then I
>> > verified that I can start JMeter from that new folder - it works
>> > fine.)
>> >
>> > Then I merged the "lib" folder from the above
>> > 'apache-jmeter-r1202266_bin.zip' file into the "lib" folder from the
>> > JMeter2.5.1_copy folder that I created.
>> >
>> > Now the JMeter in that folder won't start.
>> >
>> > I also tried merging all the files from ".zip" folder into the copy,
>> > no dice. What am I doing wrong?
>>
>> You need to unpack the _bin and _lib archives into the same directory
>> structure.
>>
>> There should be only a couple of duplicate text files at the top level.
>>
>> > --
>> > Robin D. Wilson
>> > Sr. Director of Web Development
>> > KingsIsle Entertainment, Inc.
>> > VOICE: 512-777-1861
>> > www.KingsIsle.com
>> >
>> >
>> > -----Original Message-----
>> > From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
>> > Sent: Tuesday, November 15, 2011 1:56 PM
>> > To: JMeter Users List
>> > Subject: Re: JMeter reporting higher response times
>> >
>> > You can get it here:
>> > https://builds.apache.org/job/JMeter-trunk/lastSuccessfulBuild/artif
>> > ac
>> > t/trun
>> > k/dist/
>> >
>> > Merge lib zip into bin one .
>> >
>> >
--
Cordialement.
Philippe Mouawad.
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
That doesn't work either... When I do that, the 'lib' folder is missing a
bunch of ".jar" files. The "lib" from the 'zip' file only contains 2 ".jar"
files, 'bshclient.jar' and 'jorphan.jar'.
I tried it, but that just throws a stack trace (I would show you, but I'm
pretty certain that it is related to the missing "jar" files). So I added
back the 'difference' (e.g., the "jar" files that exist in the original
JMeter2.5.1, but were missing from the JMeter2.5.1Dev after I wiped them
all, but skipping the ones that were already there from the nightly build
".zip" file). The end result was the same error as before.
Am I not getting the right nightly build file?
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
Sent: Tuesday, November 15, 2011 3:42 PM
To: JMeter Users List
Subject: Re: JMeter reporting higher response times
In your procedure, before copying jars from new zip to 2.5.1 DEV, remove all
jars from your 2.5.1 DEV. (in lib, lib/ext and bin folder).
On Tue, Nov 15, 2011 at 10:37 PM, Robin D. Wilson <rw...@gmail.com>wrote:
> I must be terribly thick headed...
>
> Here's what I did:
>
> Started with a folder called:
>
> JMeter2.5.1
>
> This contains the full (and functional) installation of JMeter 2.5.1.
> When I launch the 'bin/jmeter.bat' file from this folder, it works
> fine. The only modification to this folder is the
> 'bin/user.properties'.
>
> I copied the JMeter2.5.1 folder to JMeter2.5.1Dev. I ran the
> jmeter.bat file from this folder - it works fine.
>
> Then I extracted the 'apache-jmeter-r1202266_bin.zip' file, and copied
> the ".jar" files from the "lib" folder to the 'JMeter2.5.1Dev/lib' folder.
> I copied the "lib/etx/*.jar" files to the 'JMeter2.5.1Dev/lib/ext' folder.
> I copied the "lib/junit/test.jar" file to the 'JMeter2.5.1Dev/lib/junit'
> folder.
> I copied the "bin/*" files (not the "bin/examples" folder) to the
> 'JMeter2.5.1Dev/bin' folder.
>
> When I launch the 'bin/jmeter.bat' file, I get the following error:
>
> An error occurred: org/apache/commons/jexl2/JexlContext
> errorlevel=1
> Press any key to continue . . .
>
> (Of course, when I press a key, the cmd.exe window goes away.)
>
> I have only the JRE installed - it is possible that I need the full
> JDK in order to get this working?
>
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sebb [mailto:sebbaz@gmail.com]
> Sent: Tuesday, November 15, 2011 2:34 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> On 15 November 2011 20:28, Robin D. Wilson <rw...@gmail.com> wrote:
> > I must not be understanding what you are asking me to do...
> >
> > I downloaded the 'apache-jmeter-r1202266_bin.zip' file from the URL
> below.
> >
> > Then I 'copied' my folder for the "JMeter2.5.1" install. (Then I
> > verified that I can start JMeter from that new folder - it works
> > fine.)
> >
> > Then I merged the "lib" folder from the above
> > 'apache-jmeter-r1202266_bin.zip' file into the "lib" folder from the
> > JMeter2.5.1_copy folder that I created.
> >
> > Now the JMeter in that folder won't start.
> >
> > I also tried merging all the files from ".zip" folder into the copy,
> > no dice. What am I doing wrong?
>
> You need to unpack the _bin and _lib archives into the same directory
> structure.
>
> There should be only a couple of duplicate text files at the top level.
>
> > --
> > Robin D. Wilson
> > Sr. Director of Web Development
> > KingsIsle Entertainment, Inc.
> > VOICE: 512-777-1861
> > www.KingsIsle.com
> >
> >
> > -----Original Message-----
> > From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
> > Sent: Tuesday, November 15, 2011 1:56 PM
> > To: JMeter Users List
> > Subject: Re: JMeter reporting higher response times
> >
> > You can get it here:
> > https://builds.apache.org/job/JMeter-trunk/lastSuccessfulBuild/artif
> > ac
> > t/trun
> > k/dist/
> >
> > Merge lib zip into bin one .
> >
> > Regards
> > Philippe
> >
> > On Tue, Nov 15, 2011 at 8:46 PM, Robin D. Wilson
> > <rw...@gmail.com>
> wrote:
> >
> >> The server does use compression.
> >>
> >> I'm trying to get the nightly build now - I'll let you know the
> > difference.
> >>
> >> --
> >> Robin D. Wilson
> >> Sr. Director of Web Development
> >> KingsIsle Entertainment, Inc.
> >> VOICE: 512-777-1861
> >> www.KingsIsle.com
> >>
> >>
> >> -----Original Message-----
> >> From: sebb [mailto:sebbaz@gmail.com]
> >> Sent: Tuesday, November 15, 2011 1:27 PM
> >> To: JMeter Users List
> >> Subject: Re: JMeter reporting higher response times
> >>
> >> On 15 November 2011 19:05, Robin D. Wilson <rw...@gmail.com> wrote:
> >> > Images:
> >> >
> >> > JMeter 2.4 snap: http://tinypic.com/r/xo0hfd/5
> >> > JMeter 2.5.1 snap: http://tinypic.com/r/qn045e/5
> >>
> >> The byte sizes are very different also; larger for 2.4.
> >>
> >> That is probably due to compression - previously JMeter recorded
> >> the uncompressed size, now it records the transmitted (compressed)
size.
> >>
> >> Does the server use compression?
> >>
> >> > I'm using HTTPS.
> >> >
> >> > The POST takes longer, the GET does not. (In the images, the
> >> > "Homepage" is a GET, the "Login" is a POST.)
> >> >
> >> > Do I have to build JMeter to get the fixed HTTP client, or is
> >> > there somewhere that I can get it already ready to go (for
> >> > Windows XP, 32bit, Java JRE1.6.0_29)?
> >>
> >> Follow the nightly build links.
> >>
> >> > I consider myself a medium to advanced "user" of JMeter, but I'm
> >> > not really a developer (of any kind, anymore)...
> >> >
> >> > --
> >> > Robin D. Wilson
> >> > Sr. Director of Web Development
> >> > KingsIsle Entertainment, Inc.
> >> > VOICE: 512-777-1861
> >> > www.KingsIsle.com
> >> >
> >> >
> >> > -----Original Message-----
> >> > From: sebb [mailto:sebbaz@gmail.com]
> >> > Sent: Tuesday, November 15, 2011 11:38 AM
> >> > To: JMeter Users List
> >> > Subject: Re: JMeter reporting higher response times
> >> >
> >> > On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com>
> wrote:
> >> >> Hopefully this list can handle images...
> >> >
> >> > No, it cannot.
> >> >
> >> > Use a public hosting service and include the URL in the e-mail.
> >> >
> >> >> I have included a couple screen snaps of what I'm seeing. I ran
> >> >> a test that requests (GET) the 'homepage' from our test machine.
> >> >> Then it POSTs a login (username + password + special token) to
> >> >> the login
> >> form.
> >> >> It is a very simplified test.
> >> >>
> >> >> When I run it on JMeter 2.4, I get throughput of 128.1. When I
> >> >> run it on JMeter 2.5.1 I get throughput of 79.5.
> >> >>
> >> >> The request averages show similar differences:
> >> >>
> >> >> JMeter 2.4 - average times are 744ms
> >> >> JMeter 2.5.1 - average times are 901ms
> >> >>
> >> >> Re-running the tests over and over give very similar results
> >> >> (there's some variability in the response times of the server,
> >> >> so I had to run this a bunch of times to be sure that I wasn't
> >> >> just seeing
> > an anomaly.
> >> >> The results are consistent - each test run shows JMeter 2.4
> >> >> running significantly faster than JMeter 2.5.1.
> >> >
> >> > Do you use HTTP or HTTPS?
> >> >
> >> > Do both GET and POST show elapsed time increases?
> >> >
> >> >> BTW, just now, when I tried to switch this test to use the HTTP4
> >> >> client, the cookies stopped working correctly. So I couldn't
> >> >> test that sampler. The above numbers are just for the HTTP3.1
> >> >> client. I have a cookie value called 'stk', and using the
"${COOKIE_stk}"
> >> >> variable in the POST fails about 90% of the time when I'm using
> >> >> the
> >> >> HTTP4 client. It
> >> > works fine with HTTP3.1.
> >> >
> >> > There have been some fixes to HC4 since 2.5.1.
> >> >
> >> >> --
> >> >> Robin D. Wilson
> >> >> Sr. Director of Web Development
> >> >> KingsIsle Entertainment, Inc.
> >> >> VOICE: 512-777-1861
> >> >> www.KingsIsle.com
> >> >>
> >> >>
> >> >> -----Original Message-----
> >> >> From: sebb [mailto:sebbaz@gmail.com]
> >> >> Sent: Tuesday, November 15, 2011 10:09 AM
> >> >> To: JMeter Users List
> >> >> Subject: Re: JMeter reporting higher response times
> >> >>
> >> >> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com>
> wrote:
> >> >>> Not sure if this will matter, but have you tried testing this
> >> >>> in JMeter 2.4.x?
> >> >>>
> >> >>> The reason I ask is that I have a bunch of test scripts that I
> >> >>> regularly run in 2.4 r961953, and I have seen a significant
> >> >>> decrease in performance of these same test scripts (both
> >> >>> unmodified, and switching to various flavors of the new HTTP
> >> >>> Sampler) when I switched to
> >> >> 2.5 and 2.5.1.
> >> >>
> >> >> Can you provide details of these issues please?
> >> >>
> >> >>> --
> >> >>> Robin D. Wilson
> >> >>> Sr. Director of Web Development KingsIsle Entertainment, Inc.
> >> >>> VOICE: 512-777-1861
> >> >>> www.KingsIsle.com
> >> >>>
> >> >>>
> >> >>> -----Original Message-----
> >> >>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> >> >>> Sent: Tuesday, November 15, 2011 9:08 AM
> >> >>> To: jmeter-user@jakarta.apache.org
> >> >>> Subject: Re: JMeter reporting higher response times
> >> >>>
> >> >>> I disabled everything other than the actual sampler. Still, the
> >> >>> response time is over 2000 ms. PF attached the screenshot.
> >> >>>
> >> >>>
> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.
> >> >>> p
> >> >>> ng
> >> >>> jmeter_only_sampler.png
> >> >>>
> >> >>> Regards,
> >> >>> Sasidhar Sekar
> >> >>>
> >> >>> --
> >> >>> View this message in context:
> >> >>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-respon
> >> >>> s
> >> >>> e-
> >> >>> t
> >> >>> i
> >> >>> mes-tp
> >> >>> 4994460p4994555.html
> >> >>> Sent from the JMeter - User mailing list archive at Nabble.com.
> >> >>>
> >> >>> -----------------------------------------------------------------
> >> >>> -
> >> >>> --
> >> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >>> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>>
> >> >>>
> >> >>> -----------------------------------------------------------------
> >> >>> -
> >> >>> --
> >> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >>> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>>
> >> >>>
> >> >>
> >> >> ------------------------------------------------------------------
> >> >> -
> >> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>
> >> >>
> >> >>
> >> >> ------------------------------------------------------------------
> >> >> -
> >> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>
> >> >
> >> > -------------------------------------------------------------------
> >> > -
> >> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> > For additional commands, e-mail: user-help@jmeter.apache.org
> >> >
> >> >
> >> > -------------------------------------------------------------------
> >> > -
> >> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> > For additional commands, e-mail: user-help@jmeter.apache.org
> >> >
> >> >
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >
> >
> > --
> > Cordialement.
> > Philippe Mouawad.
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
--
Cordialement.
Philippe Mouawad.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by Philippe Mouawad <ph...@gmail.com>.
In your procedure, before copying jars from new zip to 2.5.1 DEV, remove
all jars from your 2.5.1 DEV. (in lib, lib/ext and bin folder).
On Tue, Nov 15, 2011 at 10:37 PM, Robin D. Wilson <rw...@gmail.com>wrote:
> I must be terribly thick headed...
>
> Here's what I did:
>
> Started with a folder called:
>
> JMeter2.5.1
>
> This contains the full (and functional) installation of JMeter 2.5.1. When
> I
> launch the 'bin/jmeter.bat' file from this folder, it works fine. The only
> modification to this folder is the 'bin/user.properties'.
>
> I copied the JMeter2.5.1 folder to JMeter2.5.1Dev. I ran the jmeter.bat
> file
> from this folder - it works fine.
>
> Then I extracted the 'apache-jmeter-r1202266_bin.zip' file, and copied the
> ".jar" files from the "lib" folder to the 'JMeter2.5.1Dev/lib' folder.
> I copied the "lib/etx/*.jar" files to the 'JMeter2.5.1Dev/lib/ext' folder.
> I copied the "lib/junit/test.jar" file to the 'JMeter2.5.1Dev/lib/junit'
> folder.
> I copied the "bin/*" files (not the "bin/examples" folder) to the
> 'JMeter2.5.1Dev/bin' folder.
>
> When I launch the 'bin/jmeter.bat' file, I get the following error:
>
> An error occurred: org/apache/commons/jexl2/JexlContext
> errorlevel=1
> Press any key to continue . . .
>
> (Of course, when I press a key, the cmd.exe window goes away.)
>
> I have only the JRE installed - it is possible that I need the full JDK in
> order to get this working?
>
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sebb [mailto:sebbaz@gmail.com]
> Sent: Tuesday, November 15, 2011 2:34 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> On 15 November 2011 20:28, Robin D. Wilson <rw...@gmail.com> wrote:
> > I must not be understanding what you are asking me to do...
> >
> > I downloaded the 'apache-jmeter-r1202266_bin.zip' file from the URL
> below.
> >
> > Then I 'copied' my folder for the "JMeter2.5.1" install. (Then I
> > verified that I can start JMeter from that new folder - it works
> > fine.)
> >
> > Then I merged the "lib" folder from the above
> > 'apache-jmeter-r1202266_bin.zip' file into the "lib" folder from the
> > JMeter2.5.1_copy folder that I created.
> >
> > Now the JMeter in that folder won't start.
> >
> > I also tried merging all the files from ".zip" folder into the copy,
> > no dice. What am I doing wrong?
>
> You need to unpack the _bin and _lib archives into the same directory
> structure.
>
> There should be only a couple of duplicate text files at the top level.
>
> > --
> > Robin D. Wilson
> > Sr. Director of Web Development
> > KingsIsle Entertainment, Inc.
> > VOICE: 512-777-1861
> > www.KingsIsle.com
> >
> >
> > -----Original Message-----
> > From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
> > Sent: Tuesday, November 15, 2011 1:56 PM
> > To: JMeter Users List
> > Subject: Re: JMeter reporting higher response times
> >
> > You can get it here:
> > https://builds.apache.org/job/JMeter-trunk/lastSuccessfulBuild/artifac
> > t/trun
> > k/dist/
> >
> > Merge lib zip into bin one .
> >
> > Regards
> > Philippe
> >
> > On Tue, Nov 15, 2011 at 8:46 PM, Robin D. Wilson <rw...@gmail.com>
> wrote:
> >
> >> The server does use compression.
> >>
> >> I'm trying to get the nightly build now - I'll let you know the
> > difference.
> >>
> >> --
> >> Robin D. Wilson
> >> Sr. Director of Web Development
> >> KingsIsle Entertainment, Inc.
> >> VOICE: 512-777-1861
> >> www.KingsIsle.com
> >>
> >>
> >> -----Original Message-----
> >> From: sebb [mailto:sebbaz@gmail.com]
> >> Sent: Tuesday, November 15, 2011 1:27 PM
> >> To: JMeter Users List
> >> Subject: Re: JMeter reporting higher response times
> >>
> >> On 15 November 2011 19:05, Robin D. Wilson <rw...@gmail.com> wrote:
> >> > Images:
> >> >
> >> > JMeter 2.4 snap: http://tinypic.com/r/xo0hfd/5
> >> > JMeter 2.5.1 snap: http://tinypic.com/r/qn045e/5
> >>
> >> The byte sizes are very different also; larger for 2.4.
> >>
> >> That is probably due to compression - previously JMeter recorded the
> >> uncompressed size, now it records the transmitted (compressed) size.
> >>
> >> Does the server use compression?
> >>
> >> > I'm using HTTPS.
> >> >
> >> > The POST takes longer, the GET does not. (In the images, the
> >> > "Homepage" is a GET, the "Login" is a POST.)
> >> >
> >> > Do I have to build JMeter to get the fixed HTTP client, or is there
> >> > somewhere that I can get it already ready to go (for Windows XP,
> >> > 32bit, Java JRE1.6.0_29)?
> >>
> >> Follow the nightly build links.
> >>
> >> > I consider myself a medium to advanced "user" of JMeter, but I'm
> >> > not really a developer (of any kind, anymore)...
> >> >
> >> > --
> >> > Robin D. Wilson
> >> > Sr. Director of Web Development
> >> > KingsIsle Entertainment, Inc.
> >> > VOICE: 512-777-1861
> >> > www.KingsIsle.com
> >> >
> >> >
> >> > -----Original Message-----
> >> > From: sebb [mailto:sebbaz@gmail.com]
> >> > Sent: Tuesday, November 15, 2011 11:38 AM
> >> > To: JMeter Users List
> >> > Subject: Re: JMeter reporting higher response times
> >> >
> >> > On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com>
> wrote:
> >> >> Hopefully this list can handle images...
> >> >
> >> > No, it cannot.
> >> >
> >> > Use a public hosting service and include the URL in the e-mail.
> >> >
> >> >> I have included a couple screen snaps of what I'm seeing. I ran a
> >> >> test that requests (GET) the 'homepage' from our test machine.
> >> >> Then it POSTs a login (username + password + special token) to the
> >> >> login
> >> form.
> >> >> It is a very simplified test.
> >> >>
> >> >> When I run it on JMeter 2.4, I get throughput of 128.1. When I run
> >> >> it on JMeter 2.5.1 I get throughput of 79.5.
> >> >>
> >> >> The request averages show similar differences:
> >> >>
> >> >> JMeter 2.4 - average times are 744ms
> >> >> JMeter 2.5.1 - average times are 901ms
> >> >>
> >> >> Re-running the tests over and over give very similar results
> >> >> (there's some variability in the response times of the server, so
> >> >> I had to run this a bunch of times to be sure that I wasn't just
> >> >> seeing
> > an anomaly.
> >> >> The results are consistent - each test run shows JMeter 2.4
> >> >> running significantly faster than JMeter 2.5.1.
> >> >
> >> > Do you use HTTP or HTTPS?
> >> >
> >> > Do both GET and POST show elapsed time increases?
> >> >
> >> >> BTW, just now, when I tried to switch this test to use the HTTP4
> >> >> client, the cookies stopped working correctly. So I couldn't test
> >> >> that sampler. The above numbers are just for the HTTP3.1 client. I
> >> >> have a cookie value called 'stk', and using the "${COOKIE_stk}"
> >> >> variable in the POST fails about 90% of the time when I'm using
> >> >> the
> >> >> HTTP4 client. It
> >> > works fine with HTTP3.1.
> >> >
> >> > There have been some fixes to HC4 since 2.5.1.
> >> >
> >> >> --
> >> >> Robin D. Wilson
> >> >> Sr. Director of Web Development
> >> >> KingsIsle Entertainment, Inc.
> >> >> VOICE: 512-777-1861
> >> >> www.KingsIsle.com
> >> >>
> >> >>
> >> >> -----Original Message-----
> >> >> From: sebb [mailto:sebbaz@gmail.com]
> >> >> Sent: Tuesday, November 15, 2011 10:09 AM
> >> >> To: JMeter Users List
> >> >> Subject: Re: JMeter reporting higher response times
> >> >>
> >> >> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com>
> wrote:
> >> >>> Not sure if this will matter, but have you tried testing this in
> >> >>> JMeter 2.4.x?
> >> >>>
> >> >>> The reason I ask is that I have a bunch of test scripts that I
> >> >>> regularly run in 2.4 r961953, and I have seen a significant
> >> >>> decrease in performance of these same test scripts (both
> >> >>> unmodified, and switching to various flavors of the new HTTP
> >> >>> Sampler) when I switched to
> >> >> 2.5 and 2.5.1.
> >> >>
> >> >> Can you provide details of these issues please?
> >> >>
> >> >>> --
> >> >>> Robin D. Wilson
> >> >>> Sr. Director of Web Development
> >> >>> KingsIsle Entertainment, Inc.
> >> >>> VOICE: 512-777-1861
> >> >>> www.KingsIsle.com
> >> >>>
> >> >>>
> >> >>> -----Original Message-----
> >> >>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> >> >>> Sent: Tuesday, November 15, 2011 9:08 AM
> >> >>> To: jmeter-user@jakarta.apache.org
> >> >>> Subject: Re: JMeter reporting higher response times
> >> >>>
> >> >>> I disabled everything other than the actual sampler. Still, the
> >> >>> response time is over 2000 ms. PF attached the screenshot.
> >> >>>
> >> >>>
> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.
> >> >>> p
> >> >>> ng
> >> >>> jmeter_only_sampler.png
> >> >>>
> >> >>> Regards,
> >> >>> Sasidhar Sekar
> >> >>>
> >> >>> --
> >> >>> View this message in context:
> >> >>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-respon
> >> >>> s
> >> >>> e-
> >> >>> t
> >> >>> i
> >> >>> mes-tp
> >> >>> 4994460p4994555.html
> >> >>> Sent from the JMeter - User mailing list archive at Nabble.com.
> >> >>>
> >> >>> -----------------------------------------------------------------
> >> >>> -
> >> >>> --
> >> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >>> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>>
> >> >>>
> >> >>> -----------------------------------------------------------------
> >> >>> -
> >> >>> --
> >> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >>> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>>
> >> >>>
> >> >>
> >> >> ------------------------------------------------------------------
> >> >> -
> >> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>
> >> >>
> >> >>
> >> >> ------------------------------------------------------------------
> >> >> -
> >> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >> >>
> >> >
> >> > -------------------------------------------------------------------
> >> > -
> >> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> > For additional commands, e-mail: user-help@jmeter.apache.org
> >> >
> >> >
> >> > -------------------------------------------------------------------
> >> > -
> >> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> > For additional commands, e-mail: user-help@jmeter.apache.org
> >> >
> >> >
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >
> >
> > --
> > Cordialement.
> > Philippe Mouawad.
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
--
Cordialement.
Philippe Mouawad.
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
I must be terribly thick headed...
Here's what I did:
Started with a folder called:
JMeter2.5.1
This contains the full (and functional) installation of JMeter 2.5.1. When I
launch the 'bin/jmeter.bat' file from this folder, it works fine. The only
modification to this folder is the 'bin/user.properties'.
I copied the JMeter2.5.1 folder to JMeter2.5.1Dev. I ran the jmeter.bat file
from this folder - it works fine.
Then I extracted the 'apache-jmeter-r1202266_bin.zip' file, and copied the
".jar" files from the "lib" folder to the 'JMeter2.5.1Dev/lib' folder.
I copied the "lib/etx/*.jar" files to the 'JMeter2.5.1Dev/lib/ext' folder.
I copied the "lib/junit/test.jar" file to the 'JMeter2.5.1Dev/lib/junit'
folder.
I copied the "bin/*" files (not the "bin/examples" folder) to the
'JMeter2.5.1Dev/bin' folder.
When I launch the 'bin/jmeter.bat' file, I get the following error:
An error occurred: org/apache/commons/jexl2/JexlContext
errorlevel=1
Press any key to continue . . .
(Of course, when I press a key, the cmd.exe window goes away.)
I have only the JRE installed - it is possible that I need the full JDK in
order to get this working?
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: sebb [mailto:sebbaz@gmail.com]
Sent: Tuesday, November 15, 2011 2:34 PM
To: JMeter Users List
Subject: Re: JMeter reporting higher response times
On 15 November 2011 20:28, Robin D. Wilson <rw...@gmail.com> wrote:
> I must not be understanding what you are asking me to do...
>
> I downloaded the 'apache-jmeter-r1202266_bin.zip' file from the URL below.
>
> Then I 'copied' my folder for the "JMeter2.5.1" install. (Then I
> verified that I can start JMeter from that new folder - it works
> fine.)
>
> Then I merged the "lib" folder from the above
> 'apache-jmeter-r1202266_bin.zip' file into the "lib" folder from the
> JMeter2.5.1_copy folder that I created.
>
> Now the JMeter in that folder won't start.
>
> I also tried merging all the files from ".zip" folder into the copy,
> no dice. What am I doing wrong?
You need to unpack the _bin and _lib archives into the same directory
structure.
There should be only a couple of duplicate text files at the top level.
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
> Sent: Tuesday, November 15, 2011 1:56 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> You can get it here:
> https://builds.apache.org/job/JMeter-trunk/lastSuccessfulBuild/artifac
> t/trun
> k/dist/
>
> Merge lib zip into bin one .
>
> Regards
> Philippe
>
> On Tue, Nov 15, 2011 at 8:46 PM, Robin D. Wilson <rw...@gmail.com>
wrote:
>
>> The server does use compression.
>>
>> I'm trying to get the nightly build now - I'll let you know the
> difference.
>>
>> --
>> Robin D. Wilson
>> Sr. Director of Web Development
>> KingsIsle Entertainment, Inc.
>> VOICE: 512-777-1861
>> www.KingsIsle.com
>>
>>
>> -----Original Message-----
>> From: sebb [mailto:sebbaz@gmail.com]
>> Sent: Tuesday, November 15, 2011 1:27 PM
>> To: JMeter Users List
>> Subject: Re: JMeter reporting higher response times
>>
>> On 15 November 2011 19:05, Robin D. Wilson <rw...@gmail.com> wrote:
>> > Images:
>> >
>> > JMeter 2.4 snap: http://tinypic.com/r/xo0hfd/5
>> > JMeter 2.5.1 snap: http://tinypic.com/r/qn045e/5
>>
>> The byte sizes are very different also; larger for 2.4.
>>
>> That is probably due to compression - previously JMeter recorded the
>> uncompressed size, now it records the transmitted (compressed) size.
>>
>> Does the server use compression?
>>
>> > I'm using HTTPS.
>> >
>> > The POST takes longer, the GET does not. (In the images, the
>> > "Homepage" is a GET, the "Login" is a POST.)
>> >
>> > Do I have to build JMeter to get the fixed HTTP client, or is there
>> > somewhere that I can get it already ready to go (for Windows XP,
>> > 32bit, Java JRE1.6.0_29)?
>>
>> Follow the nightly build links.
>>
>> > I consider myself a medium to advanced "user" of JMeter, but I'm
>> > not really a developer (of any kind, anymore)...
>> >
>> > --
>> > Robin D. Wilson
>> > Sr. Director of Web Development
>> > KingsIsle Entertainment, Inc.
>> > VOICE: 512-777-1861
>> > www.KingsIsle.com
>> >
>> >
>> > -----Original Message-----
>> > From: sebb [mailto:sebbaz@gmail.com]
>> > Sent: Tuesday, November 15, 2011 11:38 AM
>> > To: JMeter Users List
>> > Subject: Re: JMeter reporting higher response times
>> >
>> > On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
>> >> Hopefully this list can handle images...
>> >
>> > No, it cannot.
>> >
>> > Use a public hosting service and include the URL in the e-mail.
>> >
>> >> I have included a couple screen snaps of what I'm seeing. I ran a
>> >> test that requests (GET) the 'homepage' from our test machine.
>> >> Then it POSTs a login (username + password + special token) to the
>> >> login
>> form.
>> >> It is a very simplified test.
>> >>
>> >> When I run it on JMeter 2.4, I get throughput of 128.1. When I run
>> >> it on JMeter 2.5.1 I get throughput of 79.5.
>> >>
>> >> The request averages show similar differences:
>> >>
>> >> JMeter 2.4 - average times are 744ms
>> >> JMeter 2.5.1 - average times are 901ms
>> >>
>> >> Re-running the tests over and over give very similar results
>> >> (there's some variability in the response times of the server, so
>> >> I had to run this a bunch of times to be sure that I wasn't just
>> >> seeing
> an anomaly.
>> >> The results are consistent - each test run shows JMeter 2.4
>> >> running significantly faster than JMeter 2.5.1.
>> >
>> > Do you use HTTP or HTTPS?
>> >
>> > Do both GET and POST show elapsed time increases?
>> >
>> >> BTW, just now, when I tried to switch this test to use the HTTP4
>> >> client, the cookies stopped working correctly. So I couldn't test
>> >> that sampler. The above numbers are just for the HTTP3.1 client. I
>> >> have a cookie value called 'stk', and using the "${COOKIE_stk}"
>> >> variable in the POST fails about 90% of the time when I'm using
>> >> the
>> >> HTTP4 client. It
>> > works fine with HTTP3.1.
>> >
>> > There have been some fixes to HC4 since 2.5.1.
>> >
>> >> --
>> >> Robin D. Wilson
>> >> Sr. Director of Web Development
>> >> KingsIsle Entertainment, Inc.
>> >> VOICE: 512-777-1861
>> >> www.KingsIsle.com
>> >>
>> >>
>> >> -----Original Message-----
>> >> From: sebb [mailto:sebbaz@gmail.com]
>> >> Sent: Tuesday, November 15, 2011 10:09 AM
>> >> To: JMeter Users List
>> >> Subject: Re: JMeter reporting higher response times
>> >>
>> >> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
>> >>> Not sure if this will matter, but have you tried testing this in
>> >>> JMeter 2.4.x?
>> >>>
>> >>> The reason I ask is that I have a bunch of test scripts that I
>> >>> regularly run in 2.4 r961953, and I have seen a significant
>> >>> decrease in performance of these same test scripts (both
>> >>> unmodified, and switching to various flavors of the new HTTP
>> >>> Sampler) when I switched to
>> >> 2.5 and 2.5.1.
>> >>
>> >> Can you provide details of these issues please?
>> >>
>> >>> --
>> >>> Robin D. Wilson
>> >>> Sr. Director of Web Development
>> >>> KingsIsle Entertainment, Inc.
>> >>> VOICE: 512-777-1861
>> >>> www.KingsIsle.com
>> >>>
>> >>>
>> >>> -----Original Message-----
>> >>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
>> >>> Sent: Tuesday, November 15, 2011 9:08 AM
>> >>> To: jmeter-user@jakarta.apache.org
>> >>> Subject: Re: JMeter reporting higher response times
>> >>>
>> >>> I disabled everything other than the actual sampler. Still, the
>> >>> response time is over 2000 ms. PF attached the screenshot.
>> >>>
>> >>> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.
>> >>> p
>> >>> ng
>> >>> jmeter_only_sampler.png
>> >>>
>> >>> Regards,
>> >>> Sasidhar Sekar
>> >>>
>> >>> --
>> >>> View this message in context:
>> >>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-respon
>> >>> s
>> >>> e-
>> >>> t
>> >>> i
>> >>> mes-tp
>> >>> 4994460p4994555.html
>> >>> Sent from the JMeter - User mailing list archive at Nabble.com.
>> >>>
>> >>> -----------------------------------------------------------------
>> >>> -
>> >>> --
>> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >>> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>>
>> >>>
>> >>> -----------------------------------------------------------------
>> >>> -
>> >>> --
>> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >>> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>>
>> >>>
>> >>
>> >> ------------------------------------------------------------------
>> >> -
>> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>
>> >>
>> >>
>> >> ------------------------------------------------------------------
>> >> -
>> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>
>> >
>> > -------------------------------------------------------------------
>> > -
>> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> > For additional commands, e-mail: user-help@jmeter.apache.org
>> >
>> >
>> > -------------------------------------------------------------------
>> > -
>> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> > For additional commands, e-mail: user-help@jmeter.apache.org
>> >
>> >
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>
>
> --
> Cordialement.
> Philippe Mouawad.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by sebb <se...@gmail.com>.
On 15 November 2011 20:28, Robin D. Wilson <rw...@gmail.com> wrote:
> I must not be understanding what you are asking me to do...
>
> I downloaded the 'apache-jmeter-r1202266_bin.zip' file from the URL below.
>
> Then I 'copied' my folder for the "JMeter2.5.1" install. (Then I verified
> that I can start JMeter from that new folder - it works fine.)
>
> Then I merged the "lib" folder from the above
> 'apache-jmeter-r1202266_bin.zip' file into the "lib" folder from the
> JMeter2.5.1_copy folder that I created.
>
> Now the JMeter in that folder won't start.
>
> I also tried merging all the files from ".zip" folder into the copy, no
> dice. What am I doing wrong?
You need to unpack the _bin and _lib archives into the same directory structure.
There should be only a couple of duplicate text files at the top level.
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
> Sent: Tuesday, November 15, 2011 1:56 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> You can get it here:
> https://builds.apache.org/job/JMeter-trunk/lastSuccessfulBuild/artifact/trun
> k/dist/
>
> Merge lib zip into bin one .
>
> Regards
> Philippe
>
> On Tue, Nov 15, 2011 at 8:46 PM, Robin D. Wilson <rw...@gmail.com> wrote:
>
>> The server does use compression.
>>
>> I'm trying to get the nightly build now - I'll let you know the
> difference.
>>
>> --
>> Robin D. Wilson
>> Sr. Director of Web Development
>> KingsIsle Entertainment, Inc.
>> VOICE: 512-777-1861
>> www.KingsIsle.com
>>
>>
>> -----Original Message-----
>> From: sebb [mailto:sebbaz@gmail.com]
>> Sent: Tuesday, November 15, 2011 1:27 PM
>> To: JMeter Users List
>> Subject: Re: JMeter reporting higher response times
>>
>> On 15 November 2011 19:05, Robin D. Wilson <rw...@gmail.com> wrote:
>> > Images:
>> >
>> > JMeter 2.4 snap: http://tinypic.com/r/xo0hfd/5
>> > JMeter 2.5.1 snap: http://tinypic.com/r/qn045e/5
>>
>> The byte sizes are very different also; larger for 2.4.
>>
>> That is probably due to compression - previously JMeter recorded the
>> uncompressed size, now it records the transmitted (compressed) size.
>>
>> Does the server use compression?
>>
>> > I'm using HTTPS.
>> >
>> > The POST takes longer, the GET does not. (In the images, the
>> > "Homepage" is a GET, the "Login" is a POST.)
>> >
>> > Do I have to build JMeter to get the fixed HTTP client, or is there
>> > somewhere that I can get it already ready to go (for Windows XP,
>> > 32bit, Java JRE1.6.0_29)?
>>
>> Follow the nightly build links.
>>
>> > I consider myself a medium to advanced "user" of JMeter, but I'm not
>> > really a developer (of any kind, anymore)...
>> >
>> > --
>> > Robin D. Wilson
>> > Sr. Director of Web Development
>> > KingsIsle Entertainment, Inc.
>> > VOICE: 512-777-1861
>> > www.KingsIsle.com
>> >
>> >
>> > -----Original Message-----
>> > From: sebb [mailto:sebbaz@gmail.com]
>> > Sent: Tuesday, November 15, 2011 11:38 AM
>> > To: JMeter Users List
>> > Subject: Re: JMeter reporting higher response times
>> >
>> > On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
>> >> Hopefully this list can handle images...
>> >
>> > No, it cannot.
>> >
>> > Use a public hosting service and include the URL in the e-mail.
>> >
>> >> I have included a couple screen snaps of what I'm seeing. I ran a
>> >> test that requests (GET) the 'homepage' from our test machine. Then
>> >> it POSTs a login (username + password + special token) to the login
>> form.
>> >> It is a very simplified test.
>> >>
>> >> When I run it on JMeter 2.4, I get throughput of 128.1. When I run
>> >> it on JMeter 2.5.1 I get throughput of 79.5.
>> >>
>> >> The request averages show similar differences:
>> >>
>> >> JMeter 2.4 - average times are 744ms
>> >> JMeter 2.5.1 - average times are 901ms
>> >>
>> >> Re-running the tests over and over give very similar results
>> >> (there's some variability in the response times of the server, so I
>> >> had to run this a bunch of times to be sure that I wasn't just seeing
> an anomaly.
>> >> The results are consistent - each test run shows JMeter 2.4 running
>> >> significantly faster than JMeter 2.5.1.
>> >
>> > Do you use HTTP or HTTPS?
>> >
>> > Do both GET and POST show elapsed time increases?
>> >
>> >> BTW, just now, when I tried to switch this test to use the HTTP4
>> >> client, the cookies stopped working correctly. So I couldn't test
>> >> that sampler. The above numbers are just for the HTTP3.1 client. I
>> >> have a cookie value called 'stk', and using the "${COOKIE_stk}"
>> >> variable in the POST fails about 90% of the time when I'm using the
>> >> HTTP4 client. It
>> > works fine with HTTP3.1.
>> >
>> > There have been some fixes to HC4 since 2.5.1.
>> >
>> >> --
>> >> Robin D. Wilson
>> >> Sr. Director of Web Development
>> >> KingsIsle Entertainment, Inc.
>> >> VOICE: 512-777-1861
>> >> www.KingsIsle.com
>> >>
>> >>
>> >> -----Original Message-----
>> >> From: sebb [mailto:sebbaz@gmail.com]
>> >> Sent: Tuesday, November 15, 2011 10:09 AM
>> >> To: JMeter Users List
>> >> Subject: Re: JMeter reporting higher response times
>> >>
>> >> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
>> >>> Not sure if this will matter, but have you tried testing this in
>> >>> JMeter 2.4.x?
>> >>>
>> >>> The reason I ask is that I have a bunch of test scripts that I
>> >>> regularly run in 2.4 r961953, and I have seen a significant
>> >>> decrease in performance of these same test scripts (both
>> >>> unmodified, and switching to various flavors of the new HTTP
>> >>> Sampler) when I switched to
>> >> 2.5 and 2.5.1.
>> >>
>> >> Can you provide details of these issues please?
>> >>
>> >>> --
>> >>> Robin D. Wilson
>> >>> Sr. Director of Web Development
>> >>> KingsIsle Entertainment, Inc.
>> >>> VOICE: 512-777-1861
>> >>> www.KingsIsle.com
>> >>>
>> >>>
>> >>> -----Original Message-----
>> >>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
>> >>> Sent: Tuesday, November 15, 2011 9:08 AM
>> >>> To: jmeter-user@jakarta.apache.org
>> >>> Subject: Re: JMeter reporting higher response times
>> >>>
>> >>> I disabled everything other than the actual sampler. Still, the
>> >>> response time is over 2000 ms. PF attached the screenshot.
>> >>>
>> >>> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.
>> >>> p
>> >>> ng
>> >>> jmeter_only_sampler.png
>> >>>
>> >>> Regards,
>> >>> Sasidhar Sekar
>> >>>
>> >>> --
>> >>> View this message in context:
>> >>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-respons
>> >>> e-
>> >>> t
>> >>> i
>> >>> mes-tp
>> >>> 4994460p4994555.html
>> >>> Sent from the JMeter - User mailing list archive at Nabble.com.
>> >>>
>> >>> ------------------------------------------------------------------
>> >>> --
>> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >>> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>>
>> >>>
>> >>> ------------------------------------------------------------------
>> >>> --
>> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >>> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>>
>> >>>
>> >>
>> >> -------------------------------------------------------------------
>> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>
>> >>
>> >>
>> >> -------------------------------------------------------------------
>> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> >> For additional commands, e-mail: user-help@jmeter.apache.org
>> >>
>> >
>> > --------------------------------------------------------------------
>> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> > For additional commands, e-mail: user-help@jmeter.apache.org
>> >
>> >
>> > --------------------------------------------------------------------
>> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> > For additional commands, e-mail: user-help@jmeter.apache.org
>> >
>> >
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>
>
> --
> Cordialement.
> Philippe Mouawad.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
I must not be understanding what you are asking me to do...
I downloaded the 'apache-jmeter-r1202266_bin.zip' file from the URL below.
Then I 'copied' my folder for the "JMeter2.5.1" install. (Then I verified
that I can start JMeter from that new folder - it works fine.)
Then I merged the "lib" folder from the above
'apache-jmeter-r1202266_bin.zip' file into the "lib" folder from the
JMeter2.5.1_copy folder that I created.
Now the JMeter in that folder won't start.
I also tried merging all the files from ".zip" folder into the copy, no
dice. What am I doing wrong?
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: Philippe Mouawad [mailto:philippe.mouawad@gmail.com]
Sent: Tuesday, November 15, 2011 1:56 PM
To: JMeter Users List
Subject: Re: JMeter reporting higher response times
You can get it here:
https://builds.apache.org/job/JMeter-trunk/lastSuccessfulBuild/artifact/trun
k/dist/
Merge lib zip into bin one .
Regards
Philippe
On Tue, Nov 15, 2011 at 8:46 PM, Robin D. Wilson <rw...@gmail.com> wrote:
> The server does use compression.
>
> I'm trying to get the nightly build now - I'll let you know the
difference.
>
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sebb [mailto:sebbaz@gmail.com]
> Sent: Tuesday, November 15, 2011 1:27 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> On 15 November 2011 19:05, Robin D. Wilson <rw...@gmail.com> wrote:
> > Images:
> >
> > JMeter 2.4 snap: http://tinypic.com/r/xo0hfd/5
> > JMeter 2.5.1 snap: http://tinypic.com/r/qn045e/5
>
> The byte sizes are very different also; larger for 2.4.
>
> That is probably due to compression - previously JMeter recorded the
> uncompressed size, now it records the transmitted (compressed) size.
>
> Does the server use compression?
>
> > I'm using HTTPS.
> >
> > The POST takes longer, the GET does not. (In the images, the
> > "Homepage" is a GET, the "Login" is a POST.)
> >
> > Do I have to build JMeter to get the fixed HTTP client, or is there
> > somewhere that I can get it already ready to go (for Windows XP,
> > 32bit, Java JRE1.6.0_29)?
>
> Follow the nightly build links.
>
> > I consider myself a medium to advanced "user" of JMeter, but I'm not
> > really a developer (of any kind, anymore)...
> >
> > --
> > Robin D. Wilson
> > Sr. Director of Web Development
> > KingsIsle Entertainment, Inc.
> > VOICE: 512-777-1861
> > www.KingsIsle.com
> >
> >
> > -----Original Message-----
> > From: sebb [mailto:sebbaz@gmail.com]
> > Sent: Tuesday, November 15, 2011 11:38 AM
> > To: JMeter Users List
> > Subject: Re: JMeter reporting higher response times
> >
> > On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
> >> Hopefully this list can handle images...
> >
> > No, it cannot.
> >
> > Use a public hosting service and include the URL in the e-mail.
> >
> >> I have included a couple screen snaps of what I'm seeing. I ran a
> >> test that requests (GET) the 'homepage' from our test machine. Then
> >> it POSTs a login (username + password + special token) to the login
> form.
> >> It is a very simplified test.
> >>
> >> When I run it on JMeter 2.4, I get throughput of 128.1. When I run
> >> it on JMeter 2.5.1 I get throughput of 79.5.
> >>
> >> The request averages show similar differences:
> >>
> >> JMeter 2.4 - average times are 744ms
> >> JMeter 2.5.1 - average times are 901ms
> >>
> >> Re-running the tests over and over give very similar results
> >> (there's some variability in the response times of the server, so I
> >> had to run this a bunch of times to be sure that I wasn't just seeing
an anomaly.
> >> The results are consistent - each test run shows JMeter 2.4 running
> >> significantly faster than JMeter 2.5.1.
> >
> > Do you use HTTP or HTTPS?
> >
> > Do both GET and POST show elapsed time increases?
> >
> >> BTW, just now, when I tried to switch this test to use the HTTP4
> >> client, the cookies stopped working correctly. So I couldn't test
> >> that sampler. The above numbers are just for the HTTP3.1 client. I
> >> have a cookie value called 'stk', and using the "${COOKIE_stk}"
> >> variable in the POST fails about 90% of the time when I'm using the
> >> HTTP4 client. It
> > works fine with HTTP3.1.
> >
> > There have been some fixes to HC4 since 2.5.1.
> >
> >> --
> >> Robin D. Wilson
> >> Sr. Director of Web Development
> >> KingsIsle Entertainment, Inc.
> >> VOICE: 512-777-1861
> >> www.KingsIsle.com
> >>
> >>
> >> -----Original Message-----
> >> From: sebb [mailto:sebbaz@gmail.com]
> >> Sent: Tuesday, November 15, 2011 10:09 AM
> >> To: JMeter Users List
> >> Subject: Re: JMeter reporting higher response times
> >>
> >> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
> >>> Not sure if this will matter, but have you tried testing this in
> >>> JMeter 2.4.x?
> >>>
> >>> The reason I ask is that I have a bunch of test scripts that I
> >>> regularly run in 2.4 r961953, and I have seen a significant
> >>> decrease in performance of these same test scripts (both
> >>> unmodified, and switching to various flavors of the new HTTP
> >>> Sampler) when I switched to
> >> 2.5 and 2.5.1.
> >>
> >> Can you provide details of these issues please?
> >>
> >>> --
> >>> Robin D. Wilson
> >>> Sr. Director of Web Development
> >>> KingsIsle Entertainment, Inc.
> >>> VOICE: 512-777-1861
> >>> www.KingsIsle.com
> >>>
> >>>
> >>> -----Original Message-----
> >>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> >>> Sent: Tuesday, November 15, 2011 9:08 AM
> >>> To: jmeter-user@jakarta.apache.org
> >>> Subject: Re: JMeter reporting higher response times
> >>>
> >>> I disabled everything other than the actual sampler. Still, the
> >>> response time is over 2000 ms. PF attached the screenshot.
> >>>
> >>> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.
> >>> p
> >>> ng
> >>> jmeter_only_sampler.png
> >>>
> >>> Regards,
> >>> Sasidhar Sekar
> >>>
> >>> --
> >>> View this message in context:
> >>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-respons
> >>> e-
> >>> t
> >>> i
> >>> mes-tp
> >>> 4994460p4994555.html
> >>> Sent from the JMeter - User mailing list archive at Nabble.com.
> >>>
> >>> ------------------------------------------------------------------
> >>> --
> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >>> For additional commands, e-mail: user-help@jmeter.apache.org
> >>>
> >>>
> >>> ------------------------------------------------------------------
> >>> --
> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >>> For additional commands, e-mail: user-help@jmeter.apache.org
> >>>
> >>>
> >>
> >> -------------------------------------------------------------------
> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >>
> >> -------------------------------------------------------------------
> >> -- To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >
> > --------------------------------------------------------------------
> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
> > --------------------------------------------------------------------
> > - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
--
Cordialement.
Philippe Mouawad.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by Philippe Mouawad <ph...@gmail.com>.
You can get it here:
https://builds.apache.org/job/JMeter-trunk/lastSuccessfulBuild/artifact/trunk/dist/
Merge lib zip into bin one .
Regards
Philippe
On Tue, Nov 15, 2011 at 8:46 PM, Robin D. Wilson <rw...@gmail.com> wrote:
> The server does use compression.
>
> I'm trying to get the nightly build now - I'll let you know the difference.
>
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sebb [mailto:sebbaz@gmail.com]
> Sent: Tuesday, November 15, 2011 1:27 PM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> On 15 November 2011 19:05, Robin D. Wilson <rw...@gmail.com> wrote:
> > Images:
> >
> > JMeter 2.4 snap: http://tinypic.com/r/xo0hfd/5
> > JMeter 2.5.1 snap: http://tinypic.com/r/qn045e/5
>
> The byte sizes are very different also; larger for 2.4.
>
> That is probably due to compression - previously JMeter recorded the
> uncompressed size, now it records the transmitted (compressed) size.
>
> Does the server use compression?
>
> > I'm using HTTPS.
> >
> > The POST takes longer, the GET does not. (In the images, the
> > "Homepage" is a GET, the "Login" is a POST.)
> >
> > Do I have to build JMeter to get the fixed HTTP client, or is there
> > somewhere that I can get it already ready to go (for Windows XP,
> > 32bit, Java JRE1.6.0_29)?
>
> Follow the nightly build links.
>
> > I consider myself a medium to advanced "user" of JMeter, but I'm not
> > really a developer (of any kind, anymore)...
> >
> > --
> > Robin D. Wilson
> > Sr. Director of Web Development
> > KingsIsle Entertainment, Inc.
> > VOICE: 512-777-1861
> > www.KingsIsle.com
> >
> >
> > -----Original Message-----
> > From: sebb [mailto:sebbaz@gmail.com]
> > Sent: Tuesday, November 15, 2011 11:38 AM
> > To: JMeter Users List
> > Subject: Re: JMeter reporting higher response times
> >
> > On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
> >> Hopefully this list can handle images...
> >
> > No, it cannot.
> >
> > Use a public hosting service and include the URL in the e-mail.
> >
> >> I have included a couple screen snaps of what I'm seeing. I ran a
> >> test that requests (GET) the 'homepage' from our test machine. Then
> >> it POSTs a login (username + password + special token) to the login
> form.
> >> It is a very simplified test.
> >>
> >> When I run it on JMeter 2.4, I get throughput of 128.1. When I run it
> >> on JMeter 2.5.1 I get throughput of 79.5.
> >>
> >> The request averages show similar differences:
> >>
> >> JMeter 2.4 - average times are 744ms
> >> JMeter 2.5.1 - average times are 901ms
> >>
> >> Re-running the tests over and over give very similar results (there's
> >> some variability in the response times of the server, so I had to run
> >> this a bunch of times to be sure that I wasn't just seeing an anomaly.
> >> The results are consistent - each test run shows JMeter 2.4 running
> >> significantly faster than JMeter 2.5.1.
> >
> > Do you use HTTP or HTTPS?
> >
> > Do both GET and POST show elapsed time increases?
> >
> >> BTW, just now, when I tried to switch this test to use the HTTP4
> >> client, the cookies stopped working correctly. So I couldn't test
> >> that sampler. The above numbers are just for the HTTP3.1 client. I
> >> have a cookie value called 'stk', and using the "${COOKIE_stk}"
> >> variable in the POST fails about 90% of the time when I'm using the
> >> HTTP4 client. It
> > works fine with HTTP3.1.
> >
> > There have been some fixes to HC4 since 2.5.1.
> >
> >> --
> >> Robin D. Wilson
> >> Sr. Director of Web Development
> >> KingsIsle Entertainment, Inc.
> >> VOICE: 512-777-1861
> >> www.KingsIsle.com
> >>
> >>
> >> -----Original Message-----
> >> From: sebb [mailto:sebbaz@gmail.com]
> >> Sent: Tuesday, November 15, 2011 10:09 AM
> >> To: JMeter Users List
> >> Subject: Re: JMeter reporting higher response times
> >>
> >> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
> >>> Not sure if this will matter, but have you tried testing this in
> >>> JMeter 2.4.x?
> >>>
> >>> The reason I ask is that I have a bunch of test scripts that I
> >>> regularly run in 2.4 r961953, and I have seen a significant decrease
> >>> in performance of these same test scripts (both unmodified, and
> >>> switching to various flavors of the new HTTP Sampler) when I
> >>> switched to
> >> 2.5 and 2.5.1.
> >>
> >> Can you provide details of these issues please?
> >>
> >>> --
> >>> Robin D. Wilson
> >>> Sr. Director of Web Development
> >>> KingsIsle Entertainment, Inc.
> >>> VOICE: 512-777-1861
> >>> www.KingsIsle.com
> >>>
> >>>
> >>> -----Original Message-----
> >>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> >>> Sent: Tuesday, November 15, 2011 9:08 AM
> >>> To: jmeter-user@jakarta.apache.org
> >>> Subject: Re: JMeter reporting higher response times
> >>>
> >>> I disabled everything other than the actual sampler. Still, the
> >>> response time is over 2000 ms. PF attached the screenshot.
> >>>
> >>> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.
> >>> p
> >>> ng
> >>> jmeter_only_sampler.png
> >>>
> >>> Regards,
> >>> Sasidhar Sekar
> >>>
> >>> --
> >>> View this message in context:
> >>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-
> >>> t
> >>> i
> >>> mes-tp
> >>> 4994460p4994555.html
> >>> Sent from the JMeter - User mailing list archive at Nabble.com.
> >>>
> >>> --------------------------------------------------------------------
> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >>> For additional commands, e-mail: user-help@jmeter.apache.org
> >>>
> >>>
> >>> --------------------------------------------------------------------
> >>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >>> For additional commands, e-mail: user-help@jmeter.apache.org
> >>>
> >>>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> >> For additional commands, e-mail: user-help@jmeter.apache.org
> >>
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> > For additional commands, e-mail: user-help@jmeter.apache.org
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
--
Cordialement.
Philippe Mouawad.
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
The server does use compression.
I'm trying to get the nightly build now - I'll let you know the difference.
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: sebb [mailto:sebbaz@gmail.com]
Sent: Tuesday, November 15, 2011 1:27 PM
To: JMeter Users List
Subject: Re: JMeter reporting higher response times
On 15 November 2011 19:05, Robin D. Wilson <rw...@gmail.com> wrote:
> Images:
>
> JMeter 2.4 snap: http://tinypic.com/r/xo0hfd/5
> JMeter 2.5.1 snap: http://tinypic.com/r/qn045e/5
The byte sizes are very different also; larger for 2.4.
That is probably due to compression - previously JMeter recorded the
uncompressed size, now it records the transmitted (compressed) size.
Does the server use compression?
> I'm using HTTPS.
>
> The POST takes longer, the GET does not. (In the images, the
> "Homepage" is a GET, the "Login" is a POST.)
>
> Do I have to build JMeter to get the fixed HTTP client, or is there
> somewhere that I can get it already ready to go (for Windows XP,
> 32bit, Java JRE1.6.0_29)?
Follow the nightly build links.
> I consider myself a medium to advanced "user" of JMeter, but I'm not
> really a developer (of any kind, anymore)...
>
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sebb [mailto:sebbaz@gmail.com]
> Sent: Tuesday, November 15, 2011 11:38 AM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
>> Hopefully this list can handle images...
>
> No, it cannot.
>
> Use a public hosting service and include the URL in the e-mail.
>
>> I have included a couple screen snaps of what I'm seeing. I ran a
>> test that requests (GET) the 'homepage' from our test machine. Then
>> it POSTs a login (username + password + special token) to the login form.
>> It is a very simplified test.
>>
>> When I run it on JMeter 2.4, I get throughput of 128.1. When I run it
>> on JMeter 2.5.1 I get throughput of 79.5.
>>
>> The request averages show similar differences:
>>
>> JMeter 2.4 - average times are 744ms
>> JMeter 2.5.1 - average times are 901ms
>>
>> Re-running the tests over and over give very similar results (there's
>> some variability in the response times of the server, so I had to run
>> this a bunch of times to be sure that I wasn't just seeing an anomaly.
>> The results are consistent - each test run shows JMeter 2.4 running
>> significantly faster than JMeter 2.5.1.
>
> Do you use HTTP or HTTPS?
>
> Do both GET and POST show elapsed time increases?
>
>> BTW, just now, when I tried to switch this test to use the HTTP4
>> client, the cookies stopped working correctly. So I couldn't test
>> that sampler. The above numbers are just for the HTTP3.1 client. I
>> have a cookie value called 'stk', and using the "${COOKIE_stk}"
>> variable in the POST fails about 90% of the time when I'm using the
>> HTTP4 client. It
> works fine with HTTP3.1.
>
> There have been some fixes to HC4 since 2.5.1.
>
>> --
>> Robin D. Wilson
>> Sr. Director of Web Development
>> KingsIsle Entertainment, Inc.
>> VOICE: 512-777-1861
>> www.KingsIsle.com
>>
>>
>> -----Original Message-----
>> From: sebb [mailto:sebbaz@gmail.com]
>> Sent: Tuesday, November 15, 2011 10:09 AM
>> To: JMeter Users List
>> Subject: Re: JMeter reporting higher response times
>>
>> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
>>> Not sure if this will matter, but have you tried testing this in
>>> JMeter 2.4.x?
>>>
>>> The reason I ask is that I have a bunch of test scripts that I
>>> regularly run in 2.4 r961953, and I have seen a significant decrease
>>> in performance of these same test scripts (both unmodified, and
>>> switching to various flavors of the new HTTP Sampler) when I
>>> switched to
>> 2.5 and 2.5.1.
>>
>> Can you provide details of these issues please?
>>
>>> --
>>> Robin D. Wilson
>>> Sr. Director of Web Development
>>> KingsIsle Entertainment, Inc.
>>> VOICE: 512-777-1861
>>> www.KingsIsle.com
>>>
>>>
>>> -----Original Message-----
>>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
>>> Sent: Tuesday, November 15, 2011 9:08 AM
>>> To: jmeter-user@jakarta.apache.org
>>> Subject: Re: JMeter reporting higher response times
>>>
>>> I disabled everything other than the actual sampler. Still, the
>>> response time is over 2000 ms. PF attached the screenshot.
>>>
>>> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.
>>> p
>>> ng
>>> jmeter_only_sampler.png
>>>
>>> Regards,
>>> Sasidhar Sekar
>>>
>>> --
>>> View this message in context:
>>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-
>>> t
>>> i
>>> mes-tp
>>> 4994460p4994555.html
>>> Sent from the JMeter - User mailing list archive at Nabble.com.
>>>
>>> --------------------------------------------------------------------
>>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>>> For additional commands, e-mail: user-help@jmeter.apache.org
>>>
>>>
>>> --------------------------------------------------------------------
>>> - To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>>> For additional commands, e-mail: user-help@jmeter.apache.org
>>>
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by sebb <se...@gmail.com>.
On 15 November 2011 19:05, Robin D. Wilson <rw...@gmail.com> wrote:
> Images:
>
> JMeter 2.4 snap: http://tinypic.com/r/xo0hfd/5
> JMeter 2.5.1 snap: http://tinypic.com/r/qn045e/5
The byte sizes are very different also; larger for 2.4.
That is probably due to compression - previously JMeter recorded the
uncompressed size, now it records the transmitted (compressed) size.
Does the server use compression?
> I'm using HTTPS.
>
> The POST takes longer, the GET does not. (In the images, the "Homepage" is a
> GET, the "Login" is a POST.)
>
> Do I have to build JMeter to get the fixed HTTP client, or is there
> somewhere that I can get it already ready to go (for Windows XP, 32bit, Java
> JRE1.6.0_29)?
Follow the nightly build links.
> I consider myself a medium to advanced "user" of JMeter, but I'm not really
> a developer (of any kind, anymore)...
>
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sebb [mailto:sebbaz@gmail.com]
> Sent: Tuesday, November 15, 2011 11:38 AM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
>> Hopefully this list can handle images...
>
> No, it cannot.
>
> Use a public hosting service and include the URL in the e-mail.
>
>> I have included a couple screen snaps of what I'm seeing. I ran a test
>> that requests (GET) the 'homepage' from our test machine. Then it
>> POSTs a login (username + password + special token) to the login form.
>> It is a very simplified test.
>>
>> When I run it on JMeter 2.4, I get throughput of 128.1. When I run it
>> on JMeter 2.5.1 I get throughput of 79.5.
>>
>> The request averages show similar differences:
>>
>> JMeter 2.4 - average times are 744ms
>> JMeter 2.5.1 - average times are 901ms
>>
>> Re-running the tests over and over give very similar results (there's
>> some variability in the response times of the server, so I had to run
>> this a bunch of times to be sure that I wasn't just seeing an anomaly.
>> The results are consistent - each test run shows JMeter 2.4 running
>> significantly faster than JMeter 2.5.1.
>
> Do you use HTTP or HTTPS?
>
> Do both GET and POST show elapsed time increases?
>
>> BTW, just now, when I tried to switch this test to use the HTTP4
>> client, the cookies stopped working correctly. So I couldn't test that
>> sampler. The above numbers are just for the HTTP3.1 client. I have a
>> cookie value called 'stk', and using the "${COOKIE_stk}" variable in
>> the POST fails about 90% of the time when I'm using the HTTP4 client. It
> works fine with HTTP3.1.
>
> There have been some fixes to HC4 since 2.5.1.
>
>> --
>> Robin D. Wilson
>> Sr. Director of Web Development
>> KingsIsle Entertainment, Inc.
>> VOICE: 512-777-1861
>> www.KingsIsle.com
>>
>>
>> -----Original Message-----
>> From: sebb [mailto:sebbaz@gmail.com]
>> Sent: Tuesday, November 15, 2011 10:09 AM
>> To: JMeter Users List
>> Subject: Re: JMeter reporting higher response times
>>
>> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
>>> Not sure if this will matter, but have you tried testing this in
>>> JMeter 2.4.x?
>>>
>>> The reason I ask is that I have a bunch of test scripts that I
>>> regularly run in 2.4 r961953, and I have seen a significant decrease
>>> in performance of these same test scripts (both unmodified, and
>>> switching to various flavors of the new HTTP Sampler) when I switched
>>> to
>> 2.5 and 2.5.1.
>>
>> Can you provide details of these issues please?
>>
>>> --
>>> Robin D. Wilson
>>> Sr. Director of Web Development
>>> KingsIsle Entertainment, Inc.
>>> VOICE: 512-777-1861
>>> www.KingsIsle.com
>>>
>>>
>>> -----Original Message-----
>>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
>>> Sent: Tuesday, November 15, 2011 9:08 AM
>>> To: jmeter-user@jakarta.apache.org
>>> Subject: Re: JMeter reporting higher response times
>>>
>>> I disabled everything other than the actual sampler. Still, the
>>> response time is over 2000 ms. PF attached the screenshot.
>>>
>>> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.
>>> p
>>> ng
>>> jmeter_only_sampler.png
>>>
>>> Regards,
>>> Sasidhar Sekar
>>>
>>> --
>>> View this message in context:
>>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-t
>>> i
>>> mes-tp
>>> 4994460p4994555.html
>>> Sent from the JMeter - User mailing list archive at Nabble.com.
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>>> For additional commands, e-mail: user-help@jmeter.apache.org
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>>> For additional commands, e-mail: user-help@jmeter.apache.org
>>>
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
Images:
JMeter 2.4 snap: http://tinypic.com/r/xo0hfd/5
JMeter 2.5.1 snap: http://tinypic.com/r/qn045e/5
I'm using HTTPS.
The POST takes longer, the GET does not. (In the images, the "Homepage" is a
GET, the "Login" is a POST.)
Do I have to build JMeter to get the fixed HTTP client, or is there
somewhere that I can get it already ready to go (for Windows XP, 32bit, Java
JRE1.6.0_29)?
I consider myself a medium to advanced "user" of JMeter, but I'm not really
a developer (of any kind, anymore)...
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: sebb [mailto:sebbaz@gmail.com]
Sent: Tuesday, November 15, 2011 11:38 AM
To: JMeter Users List
Subject: Re: JMeter reporting higher response times
On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
> Hopefully this list can handle images...
No, it cannot.
Use a public hosting service and include the URL in the e-mail.
> I have included a couple screen snaps of what I'm seeing. I ran a test
> that requests (GET) the 'homepage' from our test machine. Then it
> POSTs a login (username + password + special token) to the login form.
> It is a very simplified test.
>
> When I run it on JMeter 2.4, I get throughput of 128.1. When I run it
> on JMeter 2.5.1 I get throughput of 79.5.
>
> The request averages show similar differences:
>
> JMeter 2.4 - average times are 744ms
> JMeter 2.5.1 - average times are 901ms
>
> Re-running the tests over and over give very similar results (there's
> some variability in the response times of the server, so I had to run
> this a bunch of times to be sure that I wasn't just seeing an anomaly.
> The results are consistent - each test run shows JMeter 2.4 running
> significantly faster than JMeter 2.5.1.
Do you use HTTP or HTTPS?
Do both GET and POST show elapsed time increases?
> BTW, just now, when I tried to switch this test to use the HTTP4
> client, the cookies stopped working correctly. So I couldn't test that
> sampler. The above numbers are just for the HTTP3.1 client. I have a
> cookie value called 'stk', and using the "${COOKIE_stk}" variable in
> the POST fails about 90% of the time when I'm using the HTTP4 client. It
works fine with HTTP3.1.
There have been some fixes to HC4 since 2.5.1.
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sebb [mailto:sebbaz@gmail.com]
> Sent: Tuesday, November 15, 2011 10:09 AM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
>> Not sure if this will matter, but have you tried testing this in
>> JMeter 2.4.x?
>>
>> The reason I ask is that I have a bunch of test scripts that I
>> regularly run in 2.4 r961953, and I have seen a significant decrease
>> in performance of these same test scripts (both unmodified, and
>> switching to various flavors of the new HTTP Sampler) when I switched
>> to
> 2.5 and 2.5.1.
>
> Can you provide details of these issues please?
>
>> --
>> Robin D. Wilson
>> Sr. Director of Web Development
>> KingsIsle Entertainment, Inc.
>> VOICE: 512-777-1861
>> www.KingsIsle.com
>>
>>
>> -----Original Message-----
>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
>> Sent: Tuesday, November 15, 2011 9:08 AM
>> To: jmeter-user@jakarta.apache.org
>> Subject: Re: JMeter reporting higher response times
>>
>> I disabled everything other than the actual sampler. Still, the
>> response time is over 2000 ms. PF attached the screenshot.
>>
>> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.
>> p
>> ng
>> jmeter_only_sampler.png
>>
>> Regards,
>> Sasidhar Sekar
>>
>> --
>> View this message in context:
>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-t
>> i
>> mes-tp
>> 4994460p4994555.html
>> Sent from the JMeter - User mailing list archive at Nabble.com.
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by sebb <se...@gmail.com>.
On 15 November 2011 17:24, Robin D. Wilson <rw...@gmail.com> wrote:
> Hopefully this list can handle images...
No, it cannot.
Use a public hosting service and include the URL in the e-mail.
> I have included a couple screen snaps of what I'm seeing. I ran a test that
> requests (GET) the 'homepage' from our test machine. Then it POSTs a login
> (username + password + special token) to the login form. It is a very
> simplified test.
>
> When I run it on JMeter 2.4, I get throughput of 128.1. When I run it on
> JMeter 2.5.1 I get throughput of 79.5.
>
> The request averages show similar differences:
>
> JMeter 2.4 - average times are 744ms
> JMeter 2.5.1 - average times are 901ms
>
> Re-running the tests over and over give very similar results (there's some
> variability in the response times of the server, so I had to run this a
> bunch of times to be sure that I wasn't just seeing an anomaly. The results
> are consistent - each test run shows JMeter 2.4 running significantly faster
> than JMeter 2.5.1.
Do you use HTTP or HTTPS?
Do both GET and POST show elapsed time increases?
> BTW, just now, when I tried to switch this test to use the HTTP4 client, the
> cookies stopped working correctly. So I couldn't test that sampler. The
> above numbers are just for the HTTP3.1 client. I have a cookie value called
> 'stk', and using the "${COOKIE_stk}" variable in the POST fails about 90% of
> the time when I'm using the HTTP4 client. It works fine with HTTP3.1.
There have been some fixes to HC4 since 2.5.1.
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sebb [mailto:sebbaz@gmail.com]
> Sent: Tuesday, November 15, 2011 10:09 AM
> To: JMeter Users List
> Subject: Re: JMeter reporting higher response times
>
> On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
>> Not sure if this will matter, but have you tried testing this in
>> JMeter 2.4.x?
>>
>> The reason I ask is that I have a bunch of test scripts that I
>> regularly run in 2.4 r961953, and I have seen a significant decrease
>> in performance of these same test scripts (both unmodified, and
>> switching to various flavors of the new HTTP Sampler) when I switched to
> 2.5 and 2.5.1.
>
> Can you provide details of these issues please?
>
>> --
>> Robin D. Wilson
>> Sr. Director of Web Development
>> KingsIsle Entertainment, Inc.
>> VOICE: 512-777-1861
>> www.KingsIsle.com
>>
>>
>> -----Original Message-----
>> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
>> Sent: Tuesday, November 15, 2011 9:08 AM
>> To: jmeter-user@jakarta.apache.org
>> Subject: Re: JMeter reporting higher response times
>>
>> I disabled everything other than the actual sampler. Still, the
>> response time is over 2000 ms. PF attached the screenshot.
>>
>> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.p
>> ng
>> jmeter_only_sampler.png
>>
>> Regards,
>> Sasidhar Sekar
>>
>> --
>> View this message in context:
>> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-ti
>> mes-tp
>> 4994460p4994555.html
>> Sent from the JMeter - User mailing list archive at Nabble.com.
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
>> For additional commands, e-mail: user-help@jmeter.apache.org
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
Hopefully this list can handle images...
I have included a couple screen snaps of what I'm seeing. I ran a test that
requests (GET) the 'homepage' from our test machine. Then it POSTs a login
(username + password + special token) to the login form. It is a very
simplified test.
When I run it on JMeter 2.4, I get throughput of 128.1. When I run it on
JMeter 2.5.1 I get throughput of 79.5.
The request averages show similar differences:
JMeter 2.4 - average times are 744ms
JMeter 2.5.1 - average times are 901ms
Re-running the tests over and over give very similar results (there's some
variability in the response times of the server, so I had to run this a
bunch of times to be sure that I wasn't just seeing an anomaly. The results
are consistent - each test run shows JMeter 2.4 running significantly faster
than JMeter 2.5.1.
BTW, just now, when I tried to switch this test to use the HTTP4 client, the
cookies stopped working correctly. So I couldn't test that sampler. The
above numbers are just for the HTTP3.1 client. I have a cookie value called
'stk', and using the "${COOKIE_stk}" variable in the POST fails about 90% of
the time when I'm using the HTTP4 client. It works fine with HTTP3.1.
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: sebb [mailto:sebbaz@gmail.com]
Sent: Tuesday, November 15, 2011 10:09 AM
To: JMeter Users List
Subject: Re: JMeter reporting higher response times
On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
> Not sure if this will matter, but have you tried testing this in
> JMeter 2.4.x?
>
> The reason I ask is that I have a bunch of test scripts that I
> regularly run in 2.4 r961953, and I have seen a significant decrease
> in performance of these same test scripts (both unmodified, and
> switching to various flavors of the new HTTP Sampler) when I switched to
2.5 and 2.5.1.
Can you provide details of these issues please?
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> Sent: Tuesday, November 15, 2011 9:08 AM
> To: jmeter-user@jakarta.apache.org
> Subject: Re: JMeter reporting higher response times
>
> I disabled everything other than the actual sampler. Still, the
> response time is over 2000 ms. PF attached the screenshot.
>
> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.p
> ng
> jmeter_only_sampler.png
>
> Regards,
> Sasidhar Sekar
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-ti
> mes-tp
> 4994460p4994555.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by sebb <se...@gmail.com>.
On 15 November 2011 15:58, Robin D. Wilson <rw...@gmail.com> wrote:
> Not sure if this will matter, but have you tried testing this in JMeter
> 2.4.x?
>
> The reason I ask is that I have a bunch of test scripts that I regularly run
> in 2.4 r961953, and I have seen a significant decrease in performance of
> these same test scripts (both unmodified, and switching to various flavors
> of the new HTTP Sampler) when I switched to 2.5 and 2.5.1.
Can you provide details of these issues please?
> --
> Robin D. Wilson
> Sr. Director of Web Development
> KingsIsle Entertainment, Inc.
> VOICE: 512-777-1861
> www.KingsIsle.com
>
>
> -----Original Message-----
> From: sasidharsmit [mailto:sasidharsmit@gmail.com]
> Sent: Tuesday, November 15, 2011 9:08 AM
> To: jmeter-user@jakarta.apache.org
> Subject: Re: JMeter reporting higher response times
>
> I disabled everything other than the actual sampler. Still, the response
> time is over 2000 ms. PF attached the screenshot.
>
> http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.png
> jmeter_only_sampler.png
>
> Regards,
> Sasidhar Sekar
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp
> 4994460p4994555.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
RE: JMeter reporting higher response times
Posted by "Robin D. Wilson" <rw...@gmail.com>.
Not sure if this will matter, but have you tried testing this in JMeter
2.4.x?
The reason I ask is that I have a bunch of test scripts that I regularly run
in 2.4 r961953, and I have seen a significant decrease in performance of
these same test scripts (both unmodified, and switching to various flavors
of the new HTTP Sampler) when I switched to 2.5 and 2.5.1.
If you can back-port your script to 2.4, I'd be curious if it performs any
better.
--
Robin D. Wilson
Sr. Director of Web Development
KingsIsle Entertainment, Inc.
VOICE: 512-777-1861
www.KingsIsle.com
-----Original Message-----
From: sasidharsmit [mailto:sasidharsmit@gmail.com]
Sent: Tuesday, November 15, 2011 9:08 AM
To: jmeter-user@jakarta.apache.org
Subject: Re: JMeter reporting higher response times
I disabled everything other than the actual sampler. Still, the response
time is over 2000 ms. PF attached the screenshot.
http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.png
jmeter_only_sampler.png
Regards,
Sasidhar Sekar
--
View this message in context:
http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp
4994460p4994555.html
Sent from the JMeter - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by sasidharsmit <sa...@gmail.com>.
I disabled everything other than the actual sampler. Still, the response time
is over 2000 ms. PF attached the screenshot.
http://jmeter.512774.n5.nabble.com/file/n4994555/jmeter_only_sampler.png
jmeter_only_sampler.png
Regards,
Sasidhar Sekar
--
View this message in context: http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp4994460p4994555.html
Sent from the JMeter - User mailing list archive at Nabble.com.
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
For additional commands, e-mail: user-help@jmeter.apache.org
Re: JMeter reporting higher response times
Posted by Shmuel Krakower <sh...@gmail.com>.
Try to remove the pre-processor and see how it affects the response time.
Shmuel.
On Tue, Nov 15, 2011 at 4:44 PM, sasidharsmit <sa...@gmail.com>wrote:
> Dear All,
> I'm using JMeter v2.5.1 r1177103 to automate one of my business flows, in a
> windows 7 machine. The sampler I use is a custom plugin sampler "jp@gc -
> HTTP Raw Request", that I downloaded from the below URL.
>
> http://code.google.com/p/jmeter-plugins/
>
> The business flow is simple.
> a) I send a http raw request (requesting account details for a particular
> user) and
> b) I expect a response (with the details of the user).
>
> I have completed automating the application and my script completes the
> flow
> successfully. I get the response I expect and no errors are observed.
>
> Everything is fine, except that JMeter is reporting response times higher
> than what I observe with any other performance testing tool.
>
> I have tried automating the same business flow via the following tools:
>
> 1. HTTP Requester
> 2. HP Performance Center
> 3. JMeter
>
> The response times I get via the first two tools, are similar, around 40
> ms,
> on an average. But the response times observed via JMeter is just over 2000
> ms (for every sample).
>
> I'm running the test from my local desktop for a single user, in all 3
> cases. The network capacity and the server capacity are the same in all 3
> cases.
>
> I have tried executing the tests with each of these 3 tools several times
> and yet get the same kind of result. I doubt that there is something wrong
> with the way I've designed my JMeter test plan.
>
> I have attached the screenshots of JMeter and HTTP Requester results. As
> can
> be seen in the screenshots,
>
> a) I have executed 8 tests with HTTP Requester, each returning 518 bytes
> of data, and takes around 40 ms
> b) I have executed close to 100 tests with JMeter, each returning 518
> bytes of data, and takes over 2000 ms, every time
>
> One point that interests me is the difference between "Load time" and
> "Latency" in the JMeter screenshots.
> I could see that the latency is around 10-15 ms for each of the samples but
> the load time is over 2000 ms in each sample. And, when I minus the latency
> from the load time, almost every time I get a value of around 1998-2000 ms.
>
> I do not know, exactly, what this means. But I doubt that the load time
> includes some constant time which results in such high response times.
>
> Can any of you help me find the problem area?
>
>
> http://jmeter.512774.n5.nabble.com/file/n4994460/httprequestor_screenshot.png
> httprequestor_screenshot.png
> http://jmeter.512774.n5.nabble.com/file/n4994460/jmeter_screenshot.png
> jmeter_screenshot.png
>
> Thanks and Regards,
> Sasidhar Sekar
>
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/JMeter-reporting-higher-response-times-tp4994460p4994460.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@jmeter.apache.org
> For additional commands, e-mail: user-help@jmeter.apache.org
>
>