03-10-2014 05:37 AM
I've just recorded a Web-HTTP/HTML script in URL mode (LoadRunner 11.52 build 4391).
It automatically adds web_concurrent_start(NULL); and web_concurrent_end(NULL); calls.
But it seems that everytime that a web_url or any other "web_" call is outside a concurrent "section" it fails to replay with this error:
Fatal Error -26499: Internal Error - ptVUserVars->_hDfePE is unexpectedly NOT NULL after xfeLrwSrvNetTaskItemPrimaryCreate. [MsgId: MERR-26499]
I'm new to LoadRunner, so I'm not sure if I did something wrong or if this is intended by design. But if so, then the recorder should not create such situations. Also, this error message could be a lot clearer, I found it tough to figure this out.
03-10-2014 11:18 PM
Do you have any web_reg_* steps inside the web_concurrent_group section?
LoadRunner - your performance, our passion!
03-11-2014 04:22 AM
Yes, there's some web_reg_* inside and others out.
Do you think that may be the reason? I'll to make sure none are inside and will get back here shortly.
Anyway, just to be more clear. The issue is with the calls that are outside, like exemplified by the code below:
web_concurrent_start(NULL); web_url("image.gif", "more", "parameters", LAST); web_url("print.css", "parameters", LAST); web_url("code.js", "parameters", LAST); web_concurrent_end(NULL); web_url("Example.jsp", "this one fails", LAST);
03-11-2014 05:10 AM
Ok, I moved all web_reg_* calls outside the concurrent groups. It still failed with the same error, but on a web_url call way further into my script (also outside a concurrent section). This got head-scratching... why the first calls were ok?
I notice that the call that failed was the first one outside a concurrent section after a call that had weg_regs before it, but the web_url itself was inside a concurrent section. So I also moved it out and replayed.
The script ran further and failed near the end. Same situation, a web_reg_save_param_regexp call outside a concurrent group but the matching web_submit_data inside one. I moved it outside, and the script ran successfully.
I'm not sure though which solution is preferred or easier to do:
1 - wrap all calls into concurrent sections; or
2 - remove all web_reg_ calls and its subsequent web_[url, submit_data, etc] outside concurrent sections.
Do you have any advise on those?
08-10-2014 10:00 AM - edited 08-10-2014 10:00 AM
Can someone from LR support; please conclude the topic with suitable reply to the quiestion.
You've got to jump off cliffs all the time and build your wings on the way down - Ray Bradbury
01-30-2015 05:56 AM
We were having the similar issue and found the following to be helpful and resolved the issue -
1. unselect async scan and regenerate or record the action. The replay of the newly recorded action should not receive this error.
2. don't mess with the concurrent block.
HP still need to take a look at this issue and resolve it programmitically.