| || |
1. What version of LoadRunner: v8.1.4
2. What is the protocol you are recording: web click and script
3. If HTML - are you using HTML-Advanced with URLs; N/A
4. If URL mode:n/a
5. Which LoadRunner/PerformanceCenter feature (FPs) or service packs are you using? FP1 ( I think)
6. VuGen Recording - New?
7. What kind of LoadRunner license do you have -
9. What platform(s) (PCs) and Operating Systems (Windows-XP, etc.)?
10. If you have filed a service request with HP/Mercury, what have they told you at this point with respect to your issue?
No response so far
Running a 100% load test.
From windows task manager on the injector I can see that mmdrv.exes for web click and script vusers increasing in memory ussage as test progresses( please see attached file)resulting in a fatal out of memory error..
Available physical memory reducing.
Commit charge increasing.
try with by disabling logs and avoid gui level recording.if this does not work, then have a clean load generator box,and even if this is not possible, then switch over to web protocol.
couple of other things you can give a try,
open the tasklist in the loadgenerator machine and observe which svhost process is eating the memory and kill that process,disable all the startup services which are not required etc,
how many users are running with ?normally soak test eat lot of the resources and you need to schedule proper work load with some thinktime, pacing between iteration etc depending on the reponse time you are getting on the transaction.
whatever setting you have given is correct but you need to understand that click and script does indeed consume lot of resources and most of the guys here work with web protocol as that is time tested without any issues and works much faster atleast does not throw this kind of sick errors.
i think this is what i can think max about your issue.
Hello there - thanks for the response. I wasn't aware of the C&S shortcomings. As I said I couldn't get the web scripts to iterate due to a timestamping function within the AUT. Manual correlation didn't overcome this. Anyway thanks again. It looks like I'll have to figure out a way to get the web scripts to work. Thanks.