The online community for software testing & quality assurance professionals
 
 
Calendar   Today's Topics
Sponsors:




Lost Password?

Home
BetaSoft
Blogs
Jobs
Training
News
Links
Downloads



Testing Tools >> Micro Focus/Compuware QALoad

Pages: 1
jk257
Newbie


Reged: 01/24/07
Posts: 10
QALOAD PeopleTools 8.50 - AJAX performance problem
      #627107 - 05/19/10 06:11 AM

I am wondering whether other users have experienced this 'problem' and I would like to know how they managed/solved this issue:-

Last year I ran a successful load and volume test on PeopleTools 8.49, running a load of 4500 VUs running various business processes over 1 hour 45mins. Eight player PCs were enough to manage this load.

This year I have been asked to run the test again against an upgraded PeopleTools 8.50 version. This time I am unable to run 4500 users. I can only run 20% of the load through the 8 players before their CPU goes 100% and they have memory issues. Quite a shock as I didn't expect any issues :-(

I know that PeopleTools 8.50 now uses AJAX technology. I am now resorting to using more players to help me with the load but I would like to know if anyone else has has this problem with QALOAD/AJAX/PeopleTools.

Many thanks in advance


Post Extras: Print Post   Remind Me!   Notify Moderator  
drazle
Active Member


Reged: 11/30/00
Posts: 829
Loc: Michigan, USA
Re: QALOAD PeopleTools 8.50 - AJAX performance problem [Re: jk257]
      #627322 - 05/20/10 05:02 AM

Are you using the same version of QALoad as the previous time?

--------------------
A problem is a difference between what is perceived and what is desired, that
we want to reduce
(Dewey 1933)


Post Extras: Print Post   Remind Me!   Notify Moderator  
jk257
Newbie


Reged: 01/24/07
Posts: 10
Re: QALOAD PeopleTools 8.50 - AJAX performance problem [Re: drazle]
      #627482 - 05/20/10 11:44 PM

Thanks fo replying. I have upgraded to 5.9 on QALOAD and I have IE8 installed on the machines. Intially, I used IE7, but as IE8 has Ajax support, we upgraded. IE8 works but has not made any difference.

Post Extras: Print Post   Remind Me!   Notify Moderator  
drazle
Active Member


Reged: 11/30/00
Posts: 829
Loc: Michigan, USA
Re: QALOAD PeopleTools 8.50 - AJAX performance problem [Re: jk257]
      #627542 - 05/21/10 05:45 AM

It should not matter much which browser you use when recording a script (or which is installed on player PCs). You should open a support call w/Micro Focus regarding this issue with QALoad version 5.9.

If you haven't tried recreating the script in the new version of QALoad you should try that first as that is probably the first action support will ask you to do.

--------------------
A problem is a difference between what is perceived and what is desired, that
we want to reduce
(Dewey 1933)


Post Extras: Print Post   Remind Me!   Notify Moderator  
jk257
Newbie


Reged: 01/24/07
Posts: 10
Re: QALOAD PeopleTools 8.50 - AJAX performance problem [Re: drazle]
      #629847 - 06/09/10 01:16 AM

Many thanks Drazle,
When recording I have noticed that the capture files are much bigger on the new Peopletools version. For example, the capture file for one scenario was 1.5mb for PeopleTools 4.9 but is now 16mb for peopleTools 8.5. QAload creates a script from the capture and now the script is also larger than last time - although not so drastic. I know that if the script is larger then fewer scripts will run on the PC if the PC resources remain the same. However, Capture scripts contain HTTP traffic. Will the scripts on the PCs regenerate the same amount of traffic seen in the capture script ? And will it affect the players performance. What I am trying to say is, does the script run on the players recreate the traffic load seen in the capture file? I am hoping this may be an answer to my perfomance issues. (please note that there are no issues with the environment I am load testing, it is being monitored and is coping).
I have found this statement on you site, "As a note, QALoad does not simulate a browser, we simulate user actions based on the browser used during the capture. The reason for QALoad is to generate http traffic at the protocol level. We are sending the requests that were captured to the server and parsing the replies to make sure we are getting back the expected replies." which makes me think that the capture file is important regarding player performacne but I am hoping someone can clarify the correlation btw capture file size, script size and player performance. Many thanks in advance


Post Extras: Print Post   Remind Me!   Notify Moderator  
drazle
Active Member


Reged: 11/30/00
Posts: 829
Loc: Michigan, USA
Re: QALOAD PeopleTools 8.50 - AJAX performance problem [Re: jk257]
      #630504 - 06/14/10 09:56 AM

Interesting point and essentially true I think. All settings, hardware, etc., the same with the same number of virtual users, a script that has more http requests and/or a website that has is returning larger amounts of data for the same script will take some amount more of the client resources. I'm guessing too that if it is returning more data then the reply parsing will take more client resources as well, making the biggest difference for client performance. But there are a large number of variables and some settings that can affect the PC utilization such as caching/sub-requests and java script execution level.

So you are seeing larger capture files which, if performing the exact same workflow, does indicate that the server is returning more data. Makes sense the script file isn't much larger as it only contains requests which are relatively small. I have to say that getting over 500 users per player PC is quite amazing. I've never come close to that. One thing you can try is converting in HTTP mode instead of HTML mode as HTTP mode hard-codes more sub-requests (which can be disabled) and doesn't perform java script execution. It is supposed to help with player performance. The only other thing I can think of is if you have performed any manual scripting it is possible that you may have introduced something that could hurt the performance. It is easy to introduce memory leaks with strings if you don't use the CLoadString type. Or some bad parsing, etc.

It is also possible that there may be some issue involved where QALoad is not capturing the traffic correctly for the new version of PeopleSoft. It may be worth opening a support call and sending them the relevant capture files to verify this. If Peoplesoft doesn't use propriatary encryption (such as Oracle Forms) you should consider playing back the script though a proxy capture tool (such as Fiddler ) and compare that to manually performing the workflow also captured via proxy to verify that QALoad hasn't introduced duplicated requests or something. Keep in mind that QALoad's caching doesn't really work and that it will also submit more subrequests than the browser actually does manually - that's been my experience. But this may point out a flaw if one exists.

--------------------
A problem is a difference between what is perceived and what is desired, that
we want to reduce
(Dewey 1933)


Post Extras: Print Post   Remind Me!   Notify Moderator  
jk257
Newbie


Reged: 01/24/07
Posts: 10
Re: QALOAD PeopleTools 8.50 - AJAX performance problem [Re: drazle]
      #632848 - 06/30/10 02:29 AM

Just a note to say that the business decided they wanted to test that the environment would cope with 4000+ transactions an hour. This boiled down to 583 VU'S split over 8 players (before 8.5+ Ajax). So the maximum number of vus per player was 113. Sorry to confuse over transactions and vus.

Also It has been confirmed that Web 2.0 technologies like AJAX do in fact have a higher cost for execution on client machines, which does impact how many players can reliably run, but it also requires more resources for real users, as well. Regards, JAK


Post Extras: Print Post   Remind Me!   Notify Moderator  
drazle
Active Member


Reged: 11/30/00
Posts: 829
Loc: Michigan, USA
Re: QALOAD PeopleTools 8.50 - AJAX performance problem [Re: jk257]
      #632987 - 06/30/10 12:47 PM

Interesting, but makes sense. Requires more script processing. Thanks for the follow up.

--------------------
A problem is a difference between what is perceived and what is desired, that
we want to reduce
(Dewey 1933)


Post Extras: Print Post   Remind Me!   Notify Moderator  
Pages: 1



Extra information
0 registered and 1 anonymous users are browsing this forum.

Moderator:  AJ, DSquared 

Print Topic

Forum Permissions
      You cannot start new topics
      You cannot reply to topics
      HTML is disabled
      UBBCode is enabled

Rating:
Topic views: 5958

Rate this topic

Jump to

Contact Us | Privacy statement SQAForums

Powered by UBB.threads™ 6.5.5