I am experiencing a problem in production with two specific webforms that perform a server-side postback to perform calculations.
There is a <button runat=server onserverclick=doMath>Calc</button>.
All of the data for the calculations is on the web page, and there is no database communication, but the code is written old school and everything happens server-side via postbacks; no ajax panels etc.
When the button is pressed in production, for some users, a page can not be displayed error is returned after 30-60 seconds. In the application logs on the server there is matching log entry that states an object reference was null. After testing and testing further it is clear that the data for the null reference is being sent to the web server, but it is not getting there in its entirety, and no response is making it to the user even though an error is logged.
The code seems to not be relevant, however, if that was the case, I think I would see this taking place on more than two pages. And these two pages are very similar and related to each other. However, because the problem is intermittent and it only happens to some users I also think it is a network communication problem. For example:
From home I can use the calc button over and over and I only get the error once out of 1000 clicks.
From the office I can get the error almost every single click.
The problem never takes place in dev or in qa. I am hoping for help with a method to isolate the source of the problem or maybe someone has seen this before.
EventValidation is off.
Path Pings show that there are some nodes dropping packets, but they are not "our" servers.
After cracking open Wireshark I have discovered some additional information. When the "timeout" takes place a handshake is failing.
Unfortunately, I am not a network guru. Even if this is the problem I am still concerned as it only seems to happen with two specific pages.
