|
I'm not sure if it is relevant but the following link might yield some
information. It sounds like you might have to do some tcpdump'ing on the
client-side of the director to find out what's going on though.
http://grotto11.com/blog/slash.html?+1039831658
On Friday 23 May 2008 00:24:59 JST, Benoit Gauthier wrote:
> (2008.05.22, 11:21)
>
> The following instruction was in the general definition of the ssl
> portion:
>
> BrowserMatch ".*MSIE.*" \
> nokeepalive ssl-unclean-shutdown \
> downgrade-1.0 force-response-1.0
>
> and no keepalive instruction elsewhere in the Apache config.
>
> I agree with your view that this is an IE problem, but I still need to
> get this to work!
>
> Please keep the ideas coming.
>
> Benoît
>
> > On Thu, 2008-05-22 at 10:19 -0400, Benoit Gauthier wrote:
> >> Here are observations from our most recent tests. We can replicate the
> >> problem fairly easily. The problem does NOT recur if we limit the
> >> <img> tags on the page to a single image. If we add a second <img> tag
> >> (coded as <img src=dir/image.gif>), we can produce the problem.
> >
> > Hrm. Has your Apache SSL config got the following lines?
> >
> > SetEnvIf User-Agent ".*MSIE.*" \
> > nokeepalive ssl-unclean-shutdown \
> > downgrade-1.0 force-response-1.0
> >
> > They can be set in the vhost context, or globally.
> >
> > Also, have you got keepalives switched on in your Apache config?
> >
> > This smacks to me of an app level problem being winkled out by the use
> > of LVS, but I could be wrong.
> >
> > Graeme
>
> _______________________________________________
> LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
> Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
> or go to http://lists.graemef.net/mailman/listinfo/lvs-users
|
