Full_Name: karl berry Version: 2.7.1-1.3.14 OS: solaris 2.7 Submission from: (NULL) (63.227.208.155) when connecting to a mod_ssl-enabled server with versions of internet explorer 5.00.2614.3500IC or below, going through a cisco local director, we get broken images or page-not-found errors with some frequency, on something like 1% of the connections. it is not reproducible in the sense that the same images will break every time through. it is reproducible in the sense that, sooner or later, some connection will be dropped. the problem only happens when all three of the above elements are present. if we use a later version of explorer, the bug does not happen. if we use dns round robin instead of the local director, the bug does not happen. unfortunately, we need to find a workaround that *includes* the affected explorer versions (very common), and the local director, due to circumstances beyond my control. so i am looking for a solution that just involves apache and mod_ssl. we are already doing the steps mentioned in the faq, specifically SetEnvIf User-Agent ".*MSIE.*" nokeepalive ssl-unclean-shutdown downgrade-1.0 f$ (without that, things are much worse). snooping the network interface shows that a tcp connection is opened for every image to the server, but that the connection for the broken image does not make it back out from apache. i do not know whether it gets into apache in the first place, or whether it is apache-level code or the solaris kernel that is dropping the connection. if anyone has any clues on how to determine that, as in what apache or ssl files/functions are the most likely avenues to pursue, i'd be grateful. or any other approaches to the problem, of course. thanks. ______________________________________________________________________ Apache Interface to OpenSSL (mod_ssl) www.modssl.org User Support Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
