Error -26377: No match found for the requested parameter “Siebel_Analytic_ViewState5″. Check whether the requested boundaries exist in the response data. Also, if the data you want to save exceeds 1024 bytes, use web_set_max_html_param_len to increase the parameter size [MsgId: MERR-26377]

This is a typical error on correlation but not really limited to it. This can also be a generic error for other problems and here we will discuss about the causes and how to resolve them.

So what are the possibilities of the error?

1. The server is under load and unable to return the web page correctly.

When the server is under load, it may not response causing timeout page or it may have a custom page that handles errors. In this way, a different page that was not expected by the web_reg_save_param (in Left Boundary and Right Boundary) is returned, thus causing the error. This happens usually in Controller.

2. A different request was sent prior resulting in a page that was not expected by web_reg_save_param.

Somewhere in the script before the web_reg_save_param or before the request to generate the response meant to be captured, a different request was being sent to the server which resulted in the server sending a different page. This can be caused by previously correlation (incorrect/correct ones) or paramterization.

3. The generation of the parameter of LB and RB was incorrect during recording.

The page may had changed and the LB and RB no longer exists in the application. This can be due to changes in the page design or application code.

4. The parameter within the LB and RB is more than 1024 bytes.

This occurs especially with VIEWSTATE where it will grow to a substantial size more than the initial setting of 1024 bytes. If such things, happened, increase the length with web_set_max_html_param_len accordingly to suit the application needs.

Having mentioned the above, the troubleshooting is more likely conducted using the replay log (With Full Extended Log enabled). Do not over rely on the auto-correlation, it may not be able to detect all differences in the replay and recording. The replay log is a very useful tool in tracing the requests and response during the replay. From there, you should be able to find the problem source.

I would recommend you to do manual correlation. Though tedious, you are more aware of the data sent and returned. To do this, record the application twice and compare it. Use the Wdiff tool provided by LoadRunner if you want. Also, read through the Document ID 11806 – What is correlation and how is it done.