Corrupted ViewState (Yes, another issue concerning viewstate)

[Ben Rush]

Hello World,

Okay, I have spent the day browsing the newsgroups and reading up on
article after article concerning ViewState corruption and so forth, and I
have a couple questions. We intermittantly get the following exception -

EXCEPTION MESSAGE:
The viewstate is invalid for this page and might be corrupted.

STACK TRACE:
at System.Web.UI.Page.LoadPageStateFromPersistenceMedium()
at System.Web.UI.Page.LoadPageViewState()
at System.Web.UI.Page.ProcessRequestMain()

Perhaps once or twice a day, it changes. We're NOT using a server farm,
and I have ran through the following articles, and have tried to use their
recommendations for reproducing this behavior to no avail on our machines -

http://support.microsoft.com/default.aspx?scid=kb;en-us;812399 (Not a
problem, cannot reproduce)
http://support.microsoft.com/default.aspx?scid=kb;en-us;323744 (Not a
problem, cannot reproduce)
http://support.microsoft.com/default.aspx?scid=kb;en-us;316920 (Not a
problem, cannot reproduce)
http://support.microsoft.com/default.aspx?scid=kb;en-us;324488 (Not a
problem, cannot reproduce)

The pages we seem to be getting these exceptions on have relatively
large viewstates in them, but nothing monster (with the exception of one
page which has a 125,888 bytes of ViewState). We *are* currently in the
process of shaving down the ViewState by determining which controls need to
use it and not - so size won't be nearly this issue it is now in the future.

I have encountered one thread, however at
http://groups.google.com/groups?hl=...lr=&ie=UTF-8&oe=UTF-8&scoring=d&start=10&sa=N
that peaked my interest. It's possible that the users may be timing out, and
if the validation key is autogenerated, and changes itself if the session
times out, then we could be seeing this as a potential issue.

My question is this,

1) Has anyone had any experience with this exception, and are there
possibilities for other issues outside of the support KDB articles I have
included in this mail,
2) How can I determine what the validationKey was at the time of an
exception? Is there a way I can grab this? I'm assuming not, or at least not
easily,
3) Does the validationKey timeout? And if so, how can I force it to
timeout so that I can see this issue / reproduce this issue.
4) Any other thoughts?

I understand that this issue has been beat down pretty good, but nothing
I have found so far seems to kick up any possible leads into the issue
causing this.

We're running the 1.1 framework on Server 2003.

Thanks in advance,
Ben

 

[bruce barker]

you should supply the key (same as a server farm) so that its stable between
reboots.
you should turn off EnableViewStateMac because it fails a lot.

if you done both of these the other possibility is that the user is going
thru a proxy server that truncates the viewstate. some proxy servers will
truncate a hidden field to 2k. while large viewstates are ok for entranets,
extra or internets should keep the viewstate below 1k.


-- bruce (sqlwork.com)

 

[Ben Rush]

Bruce,

I can understand the first thing you said, that is

"you should supply the key (same as a server farm) so that its stable
between reboots."

and I will make that as a recommendation. However, I would like to have a
bit more information concerning the issues being had by EnableViewStateMac,
preferrably a Microsoft KDB article, or a bit more *substance*. I'm not sure
if I like the idea of turning of the hashing of the ViewState. In other
words, when you say

"you should turn off EnableViewStateMac because it fails a lot."

Why is it failing? Or is just a unknown failure that Microsoft hasn't
recognized, etc.

Thanks,
Ben