jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Deepak Shetty <shet...@gmail.com>
Subject Re: Decrypt Hash
Date Mon, 24 Jun 2013 23:46:25 GMT
hi
again test it for multiple threads too - but few. once you are sure your
script has no issues then you can add a listener to your step 2 that only
logs responses on errors (which is still an overhead - but if you are
getting too many errors you have to discard the results of your tests
anyway).


There is no way out of storing the results for step 1) if you want
something to compare to - in some cases e.g. if your load test is returning
you error messages in the response and thats the reason why your comparison
is failing , then you might not need this




On Mon, Jun 24, 2013 at 4:41 PM, umesh prajapati <praumesh86@gmail.com>wrote:

> Thank you for the response.
> Yes the script run fine for 1 user..But, when I actually run lots of user,
> there might be some user that will have an error or may be not have error
> at all. But If it does give error for some user than I am wondering how can
> i find what the error is.
>
>
>
> On Mon, Jun 24, 2013 at 4:35 PM, Deepak Shetty <shettyd@gmail.com> wrote:
>
> > Before you try your test for a 1000 users did you try your script for a
> > single user , then a handful of users and validate your script works as
> > expected? While running this limited test you can enable saving the
> entire
> > response in addition to the hash - so you can verify what the issue is.
> >
> >
> >
> >
> > On Mon, Jun 24, 2013 at 4:13 PM, umesh prajapati <praumesh86@gmail.com
> > >wrote:
> >
> > > My scenario is like this.
> > >
> > > Since I am trying to perform a load test with big amout of users. I
> want
> > to
> > > make sure that when I run the load test it gives exact responses.
> > >
> > > Lets say i run the test for 1000 users under 5 Seconds delay. And each
> > user
> > > request will have its own responses. I save those responses. And the
> test
> > > passes with no errors.
> > > Now, I run the load test for those 1000 users I would like to compare
> or
> > > assert the responses i get during the load test exactly matches the
> > > responses I got before without load test. Or in other words I would
> like
> > to
> > > know that user1 is receiving the same response like the first test that
> > was
> > > run.
> > >
> > > To achieve this, I had to save the responses as MD5 hash and save these
> > in
> > > a file and run the load test and assert the responses exactly matches
> md5
> > > hash from previous test.
> > >
> > > Now, when I approach my test saving the responses using MD5 hash,
> > > 1) if there is error on the first test (not under load test) I am
> unable
> > to
> > > figure out what the error is.
> > > 2) if there is no error on the first test and there is error on the
> load
> > > test (which is during load test lets say user55 is not getting the
> > correct
> > > response that he is suppose to get.)
> > >
> > >
> > > On Mon, Jun 24, 2013 at 3:53 PM, Deepak Shetty <shettyd@gmail.com>
> > wrote:
> > >
> > > > hi
> > > > a) does it work even once i.e. for a single user does your script
> work?
> > > If
> > > > not - then save the entire response - debug your script whatever
> > > > b) if it fails at random (it might always be failing for the same
> user
> > or
> > > > something in which cause its same as a) ), under load then you have
> no
> > > way
> > > > out of saving the response and discarding the performance of your
> tests
> > > >
> > > > But like I said even 1 dynamic value causes your test to fail - it
> > might
> > > > not be what you are looking for.
> > > >
> > > > regards
> > > > deepak
> > > >
> > > >
> > > > On Mon, Jun 24, 2013 at 3:21 PM, umesh prajapati <
> praumesh86@gmail.com
> > > > >wrote:
> > > >
> > > > > Hi Deepak,
> > > > >
> > > > > I am running the test for very large amout of users. So, I would
> like
> > > to
> > > > > make sure that the response I am getting during the load test
> exactly
> > > > > matches the responses I am getting when I run the test with 5
> seconds
> > > > delay
> > > > > per request
> > > > >
> > > >
> > >
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message