Use multiple regexps each picking a different value in sequence, then you've
got a bunch of variables, each one unique, to do what you want with.
Or if it's a loop you could use a sequence number.
--
View this message in context:
Is there a way to improve this?
Test Plan
- Transaction Controller
- 1st Request
- 2nd Request
- 3rd Request
- 4th Request
- etc.
-Dummy Request
- Throughput Controller
My problem is if I do not use the Dummy Request and have the Throughput
Don't encode the post param (uncheck the encode checkbox next to the param in
the sampler).
Also, lose the second }, you just need one.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/HTTP-Request-tp4491018p4491052.html
Sent from the JMeter - User mailing list archive at
Total Number of Exceptions(\n|.)*\/TD
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Custom-Sampler-Tutorial-tp4490189p4492146.html
Sent from the JMeter - User mailing list archive at Nabble.com.
-
To
Ooops, you said 'TR'.
Total Number of Exceptions(\n|.)*\/TR
But you could have worked that out yourself I'm sure...
Also, weird thing to match on, why not get the actual value?
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Custom-Sampler-Tutorial-tp4490189p4492154.html
It does work. Try it http://jakarta.apache.org/oro/demo.html here or
http://rubular.com/ here . Both give correct matches.
Note, I'm only going by what you wrote - check your original post.
--
View this message in context:
Oh...you're replying to Deepak. I see.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Custom-Sampler-Tutorial-tp4490189p4492160.html
Sent from the JMeter - User mailing list archive at Nabble.com.
-
To
Hi Barrie,
I'm still new around here and can't directly answer your question but I ran
into a vaguely similar issue to the one you describe which might help you, I
found that the Shared mode (all active threads) was a bit of a non starter,
it seemed to me that using this option removed the
I think I see what's wrong here. You have 'tvuid' as your header name in the
csv file don't you?
First time I tried using the CSV Config element I hit this issue. For the
first iteration everything is fine but then on the second pass around it
sees the header row as just another line of data and
In the reg exp extractor change the radio control to use 'main samples and
sub samples' or just use a HTTP Cookie manager.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/problem-with-capturing-jessionid-with-url-redirections-tp4502249p4502364.html
Sent from the JMeter - User
I think this is a regression of Bug 49666 - CSV Header read as data after
EOF. This was marked as fixed back in 2010.
I'm a bit new at this open source business, so, tentative question: can I
reopen this bug? Do I just dive in?
--
View this message in context:
Thanks Sebb, glad I checked.
I'll go away and teach myself how the process works; I'd be happy to get
involved a bit more.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/CSV-data-file-variable-not-being-replaced-intermittently-tp4499641p4503124.html
Sent from the JMeter -
Are you looking in the response or the request?
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Using-Regular-Expression-to-extract-the-orderid-tp4506764p4507520.html
Sent from the JMeter - User mailing list archive at Nabble.com.
None of them, you can't get this metric in JMeter, only request response
time.
Std. deviation is a mathematical calculation, it has no normal value. But if
you're not setting any kind of defined pacing or regulating throughput
according to requirements then you'd possibly see something like this.
As a matter of fact you CAN add assertions that have multiple requests in
their scope.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Verify-the-response-contents-tp4509356p4509939.html
Sent from the JMeter - User mailing list archive at Nabble.com.
Much, much, more context required.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Measure-execution-time-of-a-website-tp4508895p4510899.html
Sent from the JMeter - User mailing list archive at Nabble.com.
-
Well, they do work fine, in a sense, you just have to avoid header rows. As
with any software you need to prove things work as you want them to and not
trust what it says on the tin!
But JM is, I think, the leader of the pack in the open source test tool
arena and although I often spend a bit of
200 != OK - think of this as a golden rule; It will make your life more
fulfilling.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/jMeter-http-post-requests-are-not-working-tp4424846p4512284.html
Sent from the JMeter - User mailing list archive at Nabble.com.
What happened when you experimented with other ramp up periods?
--
View this message in context:
http://jmeter.512774.n5.nabble.com/determining-ramp-up-period-tp4512401p4512425.html
Sent from the JMeter - User mailing list archive at Nabble.com.
IT nomenclature defines what you want to do as 'IP Spoofing'. A quick search
of this forum gave up the following post which provides an answer to your
question.
http://jmeter.512774.n5.nabble.com/JMeter-IP-Spoofing-td530296.html
In essence what you need to do is configure the property you listed
Hi John, When running tests like this you need to think in terms of
throughput rather than total transactions achieved. So you mention you are
aiming to get 800,000 requests but actually you want to define a throughput
rate and have this as your target. Then, the volume of requests is simply a
Just a note on the Constant Throughput business. It's not intuitive, it's
true. But if you select - 'this active thread only' and also remember that
it will apply to all threads in scope then it's actually a really solid
little workhorse - it's one of the things I like best about JM. Try putting
Try this:
Thread Group - 1 user, 1 sec ramp, forever
Dummy Request, 1 sec response time
- Constant Throughput Timer, Active Thred Only, 10 reqs per minute
Like this you get 1 request every 6 seconds, nice and steady.
Now try this:
Thread Group - 1 user, 1 sec ramp, forever
Will, The fact is, running JMeter in the Cloud is easy! In practical terms
it's no different from running it using a bunch of PCs sitting on your
desktop. But, 'The Cloud' as a phrase is somehow...exciting? It is pretty
cool, no doubt, but really, it's just hardware, hardware that is connected
Network issues...shudder! Good tip Kirk, I've only ever used ec2 and that has
outstandingly fast network connectivity - silly fast - so I can vouch for
that but it;s true virtualisation and load testing is, historically, a
touchy subject...
One more thing, today I hit an issue using a micro mysql
If the request is getting 302'd to another page and this is the request doing
the grunt work then I guess it is inside the response to this request that
you might find your id. In that case you'd need to change the config of the
Regular Expression Extractor (radio control) to Apply to: Main
The transaction controller has a little checkbox: include timer duration in
sample time (or something like that). So...
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Constant-Throughput-Timer-and-elapsed-sample-time-tp4513032p4531284.html
Sent from the JMeter - User mailing
This is your response:
msg=Invalid%20User%20ID%20/%20Password
Maybe you are passing an invalid User ID or Password?
It's a long shot, but it might be the issue.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/php-sessionid-trackin-tp4531477p4534446.html
Sent from the
It's pretty likely that the request you are making does not contain a valid
user id or password. Not guaranteed, but probable. If you actually read your
own error message you'd realise this.
This is basic stuff, you look at the request you are making a verify that it
contains the correct format /
You're not showing the actual request that is getting made, just the response
to it!
I'm bored now.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/php-sessionid-trackin-tp4531477p4534786.html
Sent from the JMeter - User mailing list archive at Nabble.com.
You may have a point, I was dismissive, but I simply stated an honest fact, I
am happy to participate here where I find the threads stimulating and where
I feel that I am giving benefit. But in this case I felt that I was no
longer giving benefit and to continue posting would be spurious.
You
Oh Man, you already said earlier that it was Felix and not Frank, sorry, you
must get that all the time!
By the way, I wasn't intending to be dismissive to you, I just felt the
original poster was getting a bit lazy and wanted to make a point...
--
View this message in context:
Unless you work for Facebook you should not be using load scripts against
their systems. Yet, a surprisingly large number of non Facebook employees
continue to try and do this nonetheless. For that reason you will quickly
find out it is difficult and, ultimately, doesn't work - once FB suspect you
Does 'almost completely 99% certain' = 98% certain?
--
View this message in context:
http://jmeter.512774.n5.nabble.com/problem-in-the-installation-of-jmeter-tp4549589p4550832.html
Sent from the JMeter - User mailing list archive at Nabble.com.
PS. Bruce is right, 100%. Just google: JAVA [Your OS].
--
View this message in context:
http://jmeter.512774.n5.nabble.com/problem-in-the-installation-of-jmeter-tp4549589p4550839.html
Sent from the JMeter - User mailing list archive at Nabble.com.
Facebook App huh, thought so. You do not need to test Facebook to test your
App, you don't want to and actually won't be able to.
You can record the requests using FB but then you should remove them and
only keep those against your own host, you might need to spoof things like
facebook_id or
Go here:
http://jakarta.apache.org/jmeter/usermanual/component_reference.html#Response_Assertion
Response Assertion
and engage eyes and head.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/How-to-use-Response-Assertion-in-Jmeter-tp4564228p4564368.html
Sent from the JMeter
Kirk is completely right. No doubt.
Virtualisation can be useful when building a test rig to use as injectors /
generators. But it is a disaster if you try and virtualise the test system
(the AUT) if your production is not also virtualised.
This is arguably worse than trying to use a scaled
Deepak, you're absolutely right, practical realities for small/medium
projects are such that this type of thing happens and always will. The trick
though, is to be aware of the limitations - if you understand why scaling
and extrapolation are flawed then you can try to mitigate.
Too many people
What leads you to think that a frame based application would behave
differently?
Also, on what basis have you excluded static resources from your test?
You may have perfectly valid reason for taking these requests out of your
test but it sounds like you are concluding that this is something
1. Due to many requests and definitely repeated requests, I thought that
it would be different from normal application.
No, I wouldn't have thought it would have much effect on the request
response traffic. As far as I know, frames are just a messy way of
displaying content.
2. All those
It's not strange, encoding would have only changed certain chars in your
string, including the + symbol. It makes sense that it worked without them.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Problem-with-authentication-in-JMeter-tp4585437p4586400.html
Sent from the
The link you posted is scary - a large proportion of what is written there is
false or misleading. If I were you, I would limit your source of information
to the help files for JMeter and avoid this site.
--
View this message in context:
Follow the steps here instead:
http:// http://jakarta.apache.org/jmeter/usermanual/index.html
http://jakarta.apache.org/jmeter/usermanual/index.html
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Parameterization-by-CSV-or-user-parameters-not-working-tp4586367p4586970.html
Would that not perhaps negate the need for a performance testing tool? Why
don't you just ask the 10,000 users to login manually and tell them to count
how long it takes and email you the results?
(By the way, this is a ridiculous idea - for one, I somehow suspect these
10,000 people have better
You're correct, not calling resources would - in most cases - be a false load
test. You need to keep in mind there are situations where leaving them out
is useful, or even correct, but typically they are something you want.
Have a look at the option 'Retrieve all Embedded Resources from HTML
There're two possible scenarios here:
1. You page is using javascript to check the browser version and this is
what gives the message.
Look for navigator.userAgent in the response / page source. If you find
this then you'll be looking at a little javascript that will also contain
the message
It's not ajax per se, it's json. Ajax is a group of technologies, more of a
concept really. Json is a format.
Chances are you need to set the content type correctly in the header; I've
noticed jmeter doesn't do this very well.
--
View this message in context:
OK, hands up, I'm pretty sure now that this is not an issue with the JMeter
Proxy (which is excellent by the way) but a problem with badboy. Bug closed.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Issue-sending-parameters-with-the-request-tp4615291p4618947.html
Sent from
Ulitimately, you want to use a database - spreadsheets are a poor solution in
this case. There's various ways to import the data and various flavours of
database to use but once there you can very easily write queries that will
compare run X with run Y, you can even trend data over time.
It all
Thing is, you seem to have the test set to run as fast as it can, without any
kind of pacing. In this situation it is entirely possible that some machines
are faster than others or that running locally vs. running in client/server
mode is also different.
But this doesn't actually matter. If you
Hi Eric, what you are talking about is running a capacity test - what level
of load can my system sustain before degradation occurs, right?
In this case you still need to define a specific load and control the rate
of requests you are sending. Sure, once you have a defind request rate there
is
Either:
1. The file is just plain not there on the remote server, did you put it
there?
2. You put it there but didn't update the path to the file in the CSV
config.
If you are trying to develop testplans on a windows machine and then run
them on a unix server, well, that sounds painful. You'd
http://rubular.com/ http://rubular.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Regular-expression-Extractor-userid-tp4658356p4662010.html
Sent from the JMeter - User mailing list archive at Nabble.com.
There's a rather fundamental point missing here. You can plan your test to
run with X threads with zero timer controls or X threads with lots of timer
controls, this might give you 1 TPS, 0.0005 TPS or 3678 TPS. But so what?
Until you actually work out what your required target load is then what's
Loops tend to iterate through rows, not columns, sounds like the 'field'
mobNo' is a column. To access different columns you can simply use a string
of comma separated variables names in the control: like:
column1,col2,mobNo,lastCol, then, you can refer to ${column1} or ${mobNo}
for each iteration
I know that it can feel like a luxury but I see it more like a mindset. It's
not that hard to get everyone together in a room and work out some high
level objectives - every project can do this, even if the best you can do is
'more than 5, less then 500'! This is very loose but it is still a
It seems like you completely understand why this request does not work, you
can see what has been recorded and you are able to work out why the request
fails: if the post params are all bunched together in one group and if you
want them to be separated into individual groups then why not just do
Use a
http://jakarta.apache.org/jmeter/usermanual/component_reference.html#Regular_Expression_Extractor
regular expression extractor
For the email:
Reference Name: foobar
Regexp: a:KeyValueOfstringguida:Key(.+?)
Template: $1$
Match No: -1
Default Value: I am the default value
This gives you:
It probably is a configuration issue - at least in the sense that the problem
only occurs under certain circumstances. You see, the thing is this
functionality normally works; in your case there is something different
which is causing a problem. Now, this may be a real code issue but if you
can
Thanks Raj, where would we be without you.
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Saving-the-reply-in-Jmeter-tp4687066p4689158.html
Sent from the JMeter - User mailing list archive at Nabble.com.
-
If you just want a random match there is a substantially easier way to do
this. It's right there in the original regular expression extractor (even
written on the form itself) just select 0 for the Match No and you will
automatically get a random Value as the result.
If you also want the count,
Kudos to Deepak, I'm going to essentially repeat what he just said with
examples, to clarify.
So, imagine this is my response data:
name=name1, value=value1
name=name2, value=value2
name=name3, value=value3
I want one regexp that will give me controlled access to each name and value
value. So
Have you tried using the HTTP Request Sampler in place of the HTTP Request
HTTPClient Sampler?
You might also want to try selecting 'Follow Redirects'.
If you google Entity enclosing requests cannot be redirected without user
intervention, there's more information for why this response is given.
The problem here is you are trying to do something that is essentially not
possible and, crucially, not useful.
JMeter is not a browser so it will never give you the exact same response
times for rendering a page as you see in your browser. Indeed, different
browsers themselves will give you
Scientifically impossible. I suppose you could try, but the universe might
implode or something.
What you are talking about is not correlation but parametisation. Big
difference. You might find reading though the basic concepts for JMeter -
http://jakarta.apache.org/jmeter/index.html here -
Still not possible. You can't extract a value from the response to a request,
and then use it in the same request - well, maybe if you had a time machine.
OK, assuming you don't have a time machine, then you can't do this.
In general, you might find benefit if you read through
Not confusing at all; you are correct on every point. You've astutely
identified that your requests (although green) are not working properly and
you've pretty much narrowed down the reason (no cookie). You barely need any
help! But you should know that in order for JM to handle cookies you need
Jreed, you are a model forum participant.
-
http://www.http503.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Unable-to-login-to-website-JMeter-noob-tp4716577p4724633.html
Sent from the JMeter - User mailing list archive at Nabble.com.
You need to http://www.google.com/search?btnG=1pws=0q=correlate+jmeter
correlate the formid.
-
http://www.http503.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Problem-with-login-tp4723687p4724673.html
Sent from the JMeter - User mailing list archive at
Sarcasm where sarcasm is due - that's what I say.
-
http://www.http503.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Unable-to-login-to-website-JMeter-noob-tp4716577p4726120.html
Sent from the JMeter - User mailing list archive at Nabble.com.
Jmeter is working fine, you simply don't have connectivity - unsurprisingly,
Jmeter needs to be able to connect in order to make a request.
Run this:
telnet [the server] 80
You need to make this command work.
-
http://www.http503.com/
--
View this message in context:
Your problem is that you do not have adequate requirements. You need to go
back to the business and get these before you can even think about what
approach to take. If you do have them but simply didn't type them here then
it is asking for problems if you ask people if you are doing the right
If you 'need' to not use a database then you have to remember that you will
be limited on how you can access the file - you will not be able to have
lots of threads trying to write to the file at the same time, this is only
really possible using a database - this is one of the reasons they were
I was dubious when I read about this Flexible File Writer but when I tested
it I discovered that it is completely true: It does write correctly from
multiple threads. Up to 500 threads in fact, each running at 1 request
(write) a second. Impressive!
OK, that makes me utterly wrong then, you do
For the thread. I ran through some tests on this and for V2.4.20110513 I get
consistent and accurate results for stdev. and avg. across all relevant
listeners.
The Test:
One thread group using a single dummy sampler. Run with two threads each
making 2 iterations.
Results:
Response Times
1. 833
Your regexp is - although not ideal - correct. (I'm making some huge
assumptions there though - you'd do better to post the relevant response
text.)
So, assuming the regexp is good, you have two scenarios:
1. You are sending the wrong ViewState - they are dynamic and will change
page to page as
Based on this response text your regexp is wrong. But somehow I suspect that
you have not provided the response text properly, at the very least try to
include the characters that follow and precede the ViewState and try
formatting it properly.
In general, point 2 answers point 1 and for point 3
Is this a new question or are you still trying to work out why your testplan
is failing?
I'm not really sure how the workings of the regular expression extractor
would relate to your problem. In short, if your regexp gives you what you
need then you should use it, it will also do some other
Trying to compare workflows in LoadRunner to how you do things in JMeter is
never going to get you anywhere. The problem is if you don't understand the
underlying concepts behind how each tools works so you will not be able to
draw any parallels and the only thing you are capable of is starting
No, no automatic creation of regexps is possible. You have to do everything
manually.
-
http://www.http503.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Issue-with-Reular-Expression-tp4770195p4774606.html
Sent from the JMeter - User mailing list archive at
You really do need to get into that View Results Tree listener. You could
post the request / response stack to this thread - detail like this is
remarkably useful.
Try using this regexp to grab the viewstate:
input\s+type=hidden\s+name=__VIEWSTATE\s+id=__VIEWSTATE\s+value=([^]+)
You might need
I've run tests like this resulting in jtl files greater than 25GB, I used the
method mentioned by Nermin, running over 20 servers each running a test that
was configured to execute only 5% of my target load. Works absolutely fine.
(Note. Amazon is your friend for such shenanigans.)
Obviously: use
Yes. HTTP Proxy and Bad Boy.
-
http://www.http503.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Can-i-use-program-to-generate-the-jmeter-script-tp4781872p4785632.html
Sent from the JMeter - User mailing list archive at Nabble.com.
I find that it is a lot simpler to Calculate Throughput based on 'This Thread
Only'. That option basically means if I set the CTT to a value of 60 and
only have 1 thread then I will get a throughput of 1 request per second. If
I have 2 threads I will get 2 requests a second, 3 threads = 3tps, etc.
Also, keep in mind. 100 threads trying to achieve 60,000 requests a second
means that each atomic process must complete within 100ms - that includes
not only the request /response but also JMeter's processing time. 100ms
isn't that long, you might find you are limited by JM or by the time your
1. You don't need to record them but you MAY need to run them. Ignore anyone
that tells you that they are not required - whether or not you should
include them is dependent on your circumstances and if you make the wrong
choice you will be running an invalid test.
2. So, yes, this option enables
Are you sure that the requests you are sending are identical to the manual
flow in the browser? I mean really the same, with the same responses?
Because if they are then you have no problem and JMeter is working fine but
if not then there really has to be a difference. Could be a header, cookie,
Yes there is, use the latest version, 2.5, this allows concurrent downloading
of embedded resources. You might also want to try clearing the browser cache
before taking timings.
But really, these sorts of comparisons are ultimately pointless - it's
better to accept that JMeter does what you tell
Also, are your manual requests getting 302'd and being redirected or are they
getting 200 responses but still redirecting?
You can see this looking at Firebug/Chrome Dev Tools/Live HTTP Headers etc.
I say this because that json string looks like it might be processed by some
javascript which
I'm thinking of running multiple jmeter processes on a single machine, up to
20. My testing so far has not highlighted any issues with this in principle
- I can execute the same jmx file over multiple java processes and I am
careful to write to separate jtl files.
The reason I'm taking this
Or better yet, think about your own particular situation:
Are you testing a public facing site or something internal?
Are you using a CDN?
Are you using content acceleration of any kind?
Where and what type of caching is in place?
At what point are you injecting load? Directly to the servers?
Thanks guys, this is really useful stuff. I'm still being impressed by what
can be done with this tool.
I'm play with this over the next few days and report back.
-
http://www.http503.com/
--
View this message in context:
What are the response times when you run these tests?
-
http://www.http503.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Constant-throughput-timer-not-giving-expected-results-tp4784904p4797538.html
Sent from the JMeter - User mailing list archive at Nabble.com.
The referer is not the request, it is from where the request was referred.
-
http://www.http503.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/How-to-pass-a-parameter-to-internal-request-embedded-tp4800759p4801913.html
Sent from the JMeter - User mailing list
Is it possible to reverse the logic in the while controller?
So, I have this:
ThreadGroup
---Sampler | This returns foo
--Regexp | My regexp is bar (thus = false)
---While | ${__javaScript(${FOO_VAR})}
--Sampler | This returns foobar
-Regexp | My
Yes, regex defaults to false. But actually, what you suggest is perfect,
thanks. For some reason I had the impression that a comparison was not
possible as the condition but it is completely logical that you should be
able to do this, it's even right there in the help.
So, for the thread
I'm not sure that you are passing things as correctly as you think. Just
ignore the referer and focus on the actual request (top of the Request tab).
These 'internal' request are not (normally) called by JMeter and you cannot
directly pass anything to them - they are variously called page
You could also tail jmeter.log.
-
http://www.http503.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/Make-JMeter-report-error-messages-in-console-tp4803852p4804129.html
Sent from the JMeter - User mailing list archive at Nabble.com.
Remember, ViewState and ViewState MAC are probably two different values
requiring two different regexps.
-
http://www.http503.com/
--
View this message in context:
http://jmeter.512774.n5.nabble.com/kingrythm-gmail-com-tp4818748p4822332.html
Sent from the JMeter - User mailing list archive
1 - 100 of 123 matches
Mail list logo
