I'm having trouble with load-balancing on a cisco content switch. The details are that I've got two servers running Apache 1.3.19 / mod_ssl/2.8.1 OpenSSL/0.9.6 and a cisco 11155 content switch (formerly known as an Arrowpoint) set up to load balance between them, and I'm browsing with MS Internet Explorer 5.5 I've got the switch set to use cookie-based balancing for HTTP. This works great. The switch is set to use the SSL session ID for HTTPS balancing. This doesn't work so well. Most of the time it works, but every once in a while, I get moved to the other server and my application breaks (since the state data is on the first server). Watching the logs and the switch itself confirms that I'm bouncing between them. I reported the problem to Cisco, and they suggested updating to a recent version of Apache, claiming that older versions regenerated the session ID too often and upgrading would fix this, but it seems that I've got fairly current versions and none of the mod_ssl changelogs mention this where I could find it. Perhaps they meant apache-ssl instead (I'm still trying to clarify this with them). While doing google searches to research the problem, I found a similar problem reported that laid the blame on IE 5 (which I'm also using), claiming that it resets the SSL connection every 90 seconds as part of an attempt to make IIS show up better in benchmarks. Does anyone have any idea what's really going on, or if either of the above statements are true? I'm using a fairly standard httpd.conf file, but I set SSLProtocol to +SSLv3 since the load balancer only supports that version. The ssl-unclean-shutdown and downgrade-1.0 settings for MSIE are still at their default values. -- [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> ______________________________________________________________________ Apache Interface to OpenSSL (mod_ssl) www.modssl.org User Support Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
