Looks like I spoke too soon. It still seems to be failing the same way as
mentioned in the original post. Any other suggestions?
"ecomputerdelicacy@yahoozier.com" wrote:
> Changing the ServicePointManager properties does seem to prevent a whole
> bunch of WebRequests from the same ServicePoint from backing up behind two
> WebRequests that are stuck. So I guess in a way that this suggestion got
> around the problem by not having back-to-back requests to the same
> ServicePoint. I still think the scenario points to a bug in the Compact
> Framework related to maximum connections per ServicePoint.
>
> In my scenario, the number of simultaneous connections is user configurable
> and may theoretically be 1000, but practically will likely stay below 20-25.
> There also may be a problem with this solution because the number of
> simultaneous connections can be increased while connections already exist.
>
> The other suggestion in the pointed to article talked about extracting the
> response from the WebException and closing the response. This is good advice,
> and I had already been following it.
>
> OVerall a good article, and I learned about the ServicePoint stuff, too.
> Thanks!
>
> I'll have to continue to run my application to see if I see the Stream mixup
> problem occurs again. I *am* still seeing "stuck" HttpWebRequest that never
> return from GetResponse. But that is another, possibly unrelated problem.
>
>
> "Daniel Moth" wrote:
>
> > See if this helps at all:
> >
http://blogs.msdn.com/davidklinems/archive/2004/09/10/228169.aspx > >
> > Cheers
> > Daniel
> > --
> >
http://www.danielmoth.com/Blog/ > >
> >
> > "ecomputerdelicacy@yahoozier.com"
> > <ecomputerdelicacyyahooziercom@discussions.microsoft.com> wrote in message
> > news:CA3B9A36-FAE5-48A3-98AC-52B1E8ED92BA@microsoft.com...
> > >I have two sequential threads that are accessing the same host using
> > > WebRequest. The first WebRequest receives a "System.Net.WebException: The
> > > operation has timed-out." The second request from the same host returns
> > > the
> > > WebRequest.GetResponseStream() from the first WebRequest URI. This seems
> > > to
> > > occur rarely (2 times in a week of several thousand requests), but
> > > regularly
> > > (every week). What's going on??!?!
> > >
> > > I'm not entirely convinced that the problem is related to threading. In
> > > one
> > > case I've seen, the second WebRequest.GetResponse() doesn't begin until
> > > well
> > > after the first thread has expired.
> > >
> > > Using Compact Framework V1.0 SP3 on WM2003SE (on a Fujitsu-Seimens Loox
> > > 720).
> > >
> > > If interested, I can point you to my application you can use to duplicate
> > > this.
> > >
> > > To email me, remove "elicacy" and "zier" from my email address.
> >