I have posted this question on two other forums in the past, and still don't know how to fix this problem. Maybe someone here knows the answer ...
As we've migrated PCs to IE8, I've noticed a steady increase in complaints about the browser timing out, usually at the most inopportune moment. The PC I'm using right now is just one of many I'm aware of in our sphere of influence that is exhibiting this problem.
If a page has to post data or get data from a server (i.e. shopping cart, payment gateway, forum post, login page), chances are the browser will time out before the operation completes. Whatever the user was doing will be replaced by a “The page cannot be displayed” error. In the case of a post operation, there is usually no way to tell whether the operation completed (the order was placed, the message was posted, etc.) This just happened to me here on the Microsoft Forums while trying to post an earlier message.
I have seen this happen on a multitude of PCs of all shapes and sizes, mostly (but not always) running Windows XP, on about a dozen different networks (P2P, Win2003, SBS2003) scattered throughout town.
This problem does not happen AT ALL – on the same PCs, the same sites, the same pages – if the user switches browsers. So, in addition to delaying the implementation of IE8 as much as possible, we’ve taken to installing Firefox on all PCs we manage.
The last time I posted this question on an MS-managed board, it was treated as if I were having an isolated problem on one screwy PC. But this is not an isolated problem. This is not one screwy PC. It is a growing number of PCs who only become screwy after IE8 is installed.
How do I solve this?
If a page has to post data or get data from a server (i.e. shopping cart, payment gateway, forum post, login page), chances are the browser will time out before the operation completes.
Since those examples probably involve HTTPS there isn't much that you can do in the way of diagnostics. E.g. you could try using Fiddler2 but as explained in its documentation that creates a "man-in-the-middle" scenario which makes certificates invalid and thus changes both the timing and the UI.
So the only remaining diagnostics that I'm aware of are less specific about your symptom and only provide timing information about secondary details, e.g. use WireShark or NetMon for packet tracing and ProcMon for file and registry accesses and task switching.
I shouldn't have to trace packets. This problem begins on any given PC the moment IE8 is installed. The user had no problem with a particular Web site (or sites) prior to IE8, and after IE8 is installed he has repeated issues with that same site site (or sites). And switching to Firefox cures the problem 100% of the time. In fact, performing that identical operation in Firefox immediately after it has failed in IE8 works in Firefox 100% of the time, which is causing many of my users to switch permanerntly to Firefox.
This is an issue with some hair-trigger timeout value in IE8 somewhere. The question is, where?
In fact, performing that identical operation in Firefox immediately after it has failed in IE8 works in Firefox 100% of the time, which is causing many of my users to switch permanerntly to Firefox.
All the more reason to trace, for comparison purposes. However, it would be more convincing to see Firefox' trace without having IE prep the environment for it. <eg>