You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@jmeter.apache.org by Takashi Imoto <t-...@ak.jp.nec.com> on 2019/11/27 11:51:17 UTC
RE: [OpenIV:0657] Re: Bugzilla Enhancement: "Auto-Correlation as
core functionality in JMeter"
To: Scott-san and Vladimir-san
Scott-san thank you for explain instead of me.
Your explanation is perfectly true !
One thing, I want to add explain and please give me idea.
> Two Scripts: Record the same sequence of actions twice then run a
> comparison of the two recordings (.jmx files) and look for where data and
> parameters differ.
As Scott-san explain, to find correlation variable, we compare 2 JMX files.
But only JMX file, it lack information to replace parameter.
To replace parameter, JMeter use CSS/JQuery extractor and it need to set location of variable.
Example.
In HTML file,
<id="javax.faces.ViewState" value=.....>
At CSS extractor
To set following.
id=[$1]
These location information are only written in JTL file not JMX file, so it need to load JTL file.
Instead of this, after recording using HTTP Test Script Recorder¶, these information is saved in memory.
So after recording and load previous recoded JMX file, it can find correlation variable and replace by CSS/JQuery extractor.
That's say.
" Test Plan must be recorded using the documentation shared here."
Is this limitation acceptable? Or is there any approach?
> -----Original Message-----
> From: Scott Dean <ps...@ncsu.edu>
> Sent: Wednesday, November 27, 2019 3:47 AM
> To: dev@jmeter.apache.org
> Cc: Rahul Mehta <rm...@gmail.com>; necti_openiv_swe
> <ne...@swe.jp.nec.com>
> Subject: [OpenIV:0657] Re: Bugzilla Enhancement: "Auto-Correlation as core
> functionality in JMeter"
>
> Greetings!
>
> I am not part of this project to add auto-correlation to JMeter but here is my take
> on it:
>
> (Note: I use the words "script" and "test plan" interchangeably)
>
> Correlation is the process of capturing the session-specific information passed
> from server to client and back to server for purposes of maintaining state, etc in
> http interactions.
>
> Manual correlation is where the person(s) developing the JMeter test plan
> manually identify the parameter(s) that needs to be captured, find where the
> values of those parameters were initially introduced in the script, inserts the
> appropriate extractors to capture the values, then replaces the recorded values
> with the JMeter variables used to hold the extracted values.
>
>
> Auto-correlation is doing this process programmatically.
>
> In general, auto-correlation is done one of two ways:
>
> * Two Scripts: Record the same sequence of actions twice then run a
> comparison of the two recordings (.jmx files) and look for where data and
> parameters differ. Ignoring differences that are generally irrelevant, like
> timestamps, present those differences to the JMeter user, allowing each
> difference to be accepted or ignored by the user. For accepted differences,
> automatically insert a relevant extractor at the relevant place and
> search/replace the recorded value with the JMeter variable used in relevant
> extraction. These steps are performed on one of the recorded scripts,
> generally the initial recorded script; the other recorded script is generally
> discarded after the process is complete.
>
>
> * Rules: Using rules that define the names of parameters that need to be
> captured and how to capture them, these rules are applied to a test plan.
> Based on criteria specified in the rule, the test plan is analyzed, rule matches
> identified, extractors insert, and hard-coded parameter values replaced with
> JMeter variables holding the extracted values
>
> I will stop here.
>
> As I said, this is my understanding of the process and may or may not be exactly
> the processes/methods used in the proposed enhancement. Feel free to
> correct me.
>
> Hope that helps!
>
> -- Scott Dean
> North Carolina State University
> Raleigh, NC USA
>
> On Tue, Nov 26, 2019 at 5:53 AM Vladimir Sitnikov
> <sitnikov.vladimir@gmail.com <ma...@gmail.com> > wrote:
>
>
> >About new feature of auto-correlation, we prepared specification
> written
> as manual.
>
> That is great. It really helps to preview the feature.
>
> However, I fail to understand what is "correlation with JMX file".
> I see it consumes an extra JMX file, but it is not clear what should be
> there in the file,
> it is not clear what is added as a result.
>
> Does "Correlation" differ from "Auto-Correlation"? Should just one of
> them
> be used in all the places?
> Is there a "manual correlation"?
>
> How Rule differs from Correlation?
> Is Rule mandatory for Correlation to work?
>
> Vladimir
>
RE: [OpenIV:0551] Re: Bugzilla Enhancement: "Auto-Correlation as core
functionality in JMeter"
Posted by Rahul Mehta <rm...@gmail.com>.
Hi Imoto,
Please clear below points
On Thursday, December 12, 2019, Takashi Imoto <t-...@ak.jp.nec.com> wrote:
> To: Valdimir-san.
>
> > Does "Correlation" differ from "Auto-Correlation"? Should just one of
> them be
> > used in all the places? Is there a "manual correlation"?
>
> "Auto-Correlation" mean, list up candidate of correlation variables.
> From these variable, user select correlate variable.
>
> Why you are manually selecting the corealtion candidates? it should be
automated. Why are you going for table based approach, any specific
reason??
> How Rule differs from Correlation?
>
> Rule files use for re-use of correlation.
> I think same application and framework correlation variable is same.
> So 1st create correlation rule that describe which variable correlate, the
> load this file and script changed.
>
Rule should be applied once for all apis where we have correlation and
accordingly script should be updated.
> > Is Rule mandatory for Correlation to work?
>
> As describe above, rule file is for re-use and this file is option.
>
> Is my explanation is clear?
> If there are other question and concern, please add comments.
I don't think any significance of rule file here, if you want to reuse
corelation then candidates should be picked from existing apis by reading
in memory, any specific reason??
>
> Thanks.
>
> > -----Original Message-----
> > From: Vladimir Sitnikov <si...@gmail.com>
> > Sent: Tuesday, November 26, 2019 7:53 PM
> > To: ApacheJMeter dev list <de...@jmeter.apache.org>
> > Cc: Rahul Mehta <rm...@gmail.com>; necti_openiv_swe
> > <ne...@swe.jp.nec.com>
> > Subject: [OpenIV:0656] Re: Bugzilla Enhancement: "Auto-Correlation as
> core
> > functionality in JMeter"
> >
> > >About new feature of auto-correlation, we prepared specification
> written as
> > manual.
> >
> > That is great. It really helps to preview the feature.
> >
> > However, I fail to understand what is "correlation with JMX file".
> > I see it consumes an extra JMX file, but it is not clear what should be
> there in
> > the file, it is not clear what is added as a result.
> >
> > Does "Correlation" differ from "Auto-Correlation"? Should just one of
> them be
> > used in all the places?
> > Is there a "manual correlation"?
> >
> > How Rule differs from Correlation?
> >
> > Is Rule mandatory for Correlation to work?
> >
> > Vladimir
>
>
RE: [OpenIV:0551] Re: Bugzilla Enhancement: "Auto-Correlation as
core functionality in JMeter"
Posted by Takashi Imoto <t-...@ak.jp.nec.com>.
To Rahul-san.
> I have same understanding. I am unable to understand that why extra jmx file is
> used, it is not clear from the file as outcome.
To list up correlate variable, we need to compare 2 jmx file and find and list up variable that value is different.
So it need extra jmx file.
> Imoto,
> Can we have better way rather than selecting correlation candidates manually in table?
Table show candidate of variable and value.
I think from this information, user easily select which variable are correlate value.
(If value is random, it might be correlate variable)
To avoid miss setting, I add manual selected process.
I thinks other tools do same process.
> Auto-correlation should not involve manual steps and why rules are mandatory here.
Rule file are not mandatory, it is option file that describe previous mail.
I think it can skip manual selecting, but to avoid miss setting and to confirm, it might be better to show table and select correlate variable.
> I have seen same implementation in some other load testing tool, so it would be
> great if you can share patent list.
I check LoadRunner’s patent list as following, but these patent does not refer to correlating.
5,701,139 System for tracking and replicating the operation of a cursor manipulation device
5,657,438 Interactive system for developing tests of system under test allowing independent positioning of execution start and stop markers to execute subportion of test script
5,511,185 System for automatic testing of computer software having output synchronization and capable of responding to asynchronous events
5,870,559 Software system and associated methods for facilitating the analysis and management of web sites
5,958,008 Software system and associated methods for scanning and mapping dynamically-generated web documents
5,974,572 Software system and methods for generating a load test using a server access log
6,138,157 Method and apparatus for testing web sites
6,144,962 Visualization of web sites and hierarchical data structures
6,205,122 Automatic network topology analysis
6,237,006 Methods for graphically representing web sites and hierarchical node structures
6,341,310 System and methods for facilitating the viewing and analysis of web site usage data
6,360,332 Software system and methods for testing the functionality of a transactional server
Thanks.
> -----Original Message-----
> From: Rahul Mehta <rm...@gmail.com>
> Sent: Tuesday, December 10, 2019 4:16 AM
> To: Vladimir Sitnikov <si...@gmail.com>
> Cc: ApacheJMeter dev list <de...@jmeter.apache.org>; necti_openiv_swe
> <ne...@swe.jp.nec.com>
> Subject: RE: [OpenIV:0551] Re: Bugzilla Enhancement: "Auto-Correlation as
> core functionality in JMeter"
>
> Hi Vladimir,
>
> I have same understanding. I am unable to understand that why extra jmx file is
> used, it is not clear from the file as outcome.
>
> Imoto,
> Can we have better way rather than selecting correlation candidates manually
> in table?
>
> Auto-correlation should not involve manual steps and why rules are mandatory
> here.
>
> I have seen same implementation in some other load testing tool, so it would be
> great if you can share patent list.
>
>
> Regards,
> Rahul
>
> On Tuesday, November 26, 2019, Vladimir Sitnikov <
> sitnikov.vladimir@gmail.com<ma...@gmail.com>> wrote:
>
> > >About new feature of auto-correlation, we prepared specification
> > >written
> > as manual.
> >
> > That is great. It really helps to preview the feature.
> >
> > However, I fail to understand what is "correlation with JMX file".
> > I see it consumes an extra JMX file, but it is not clear what should
> > be there in the file, it is not clear what is added as a result.
> >
> > Does "Correlation" differ from "Auto-Correlation"? Should just one of
> > them be used in all the places?
> > Is there a "manual correlation"?
> >
> > How Rule differs from Correlation?
> > Is Rule mandatory for Correlation to work?
> >
> > Vladimir
> >
> >
RE: [OpenIV:0551] Re: Bugzilla Enhancement: "Auto-Correlation as core
functionality in JMeter"
Posted by Rahul Mehta <rm...@gmail.com>.
Hi Vladimir,
I have same understanding. I am unable to understand that why extra jmx
file is used, it is not clear from the file as outcome.
Imoto,
Can we have better way rather than selecting correlation candidates
manually in table?
Auto-correlation should not involve manual steps and why rules are
mandatory here.
I have seen same implementation in some other load testing tool, so it
would be great if you can share patent list.
Regards,
Rahul
On Tuesday, November 26, 2019, Vladimir Sitnikov <
sitnikov.vladimir@gmail.com> wrote:
> >About new feature of auto-correlation, we prepared specification written
> as manual.
>
> That is great. It really helps to preview the feature.
>
> However, I fail to understand what is "correlation with JMX file".
> I see it consumes an extra JMX file, but it is not clear what should be
> there in the file,
> it is not clear what is added as a result.
>
> Does "Correlation" differ from "Auto-Correlation"? Should just one of them
> be used in all the places?
> Is there a "manual correlation"?
>
> How Rule differs from Correlation?
> Is Rule mandatory for Correlation to work?
>
> Vladimir
>
>