Jeffrey Altman wrote: >Since absolutely no one that is experiencing this problem >has looked at a suffering process in a debugger it is >impossible to know what is the cause of the problem. As >far as I can tell all the theories that have been put >forward as to what this is or is not are simply best guesses >without much evidence to back them up.
We've diagnosed a particular behavior with Apache 1.3.26 on the OpenVMS platform where: a) all available HTTPS connections are open to the same client address and the clients do not start or complete the SSL handshake b) these initializing HTTPS connections do not timeout (a) is interesting from an Apache/mod_ssl/OpenSSL point of view, while (b) is a platform-specific problem related to SIGALRM signals and blocking recv() calls that may or may not affect other platforms. (a) is caused by client software initiating an HTTPS connection and never sending data across the connection. The server waits for data to be read on the connection in SSL_accept(). When a timeout signal (SIGALRM) is received, the recv() completes with -1 and mod_ssl shuts down the connection. This is expected behavior for mod_ssl and OpenSSL and is designed to prevent "hung" connections from being kept open by unresponsive clients (see (b)). mod_ssl\ssl_engine_kernel\ssl_hook_NewConnection: ______________________________________________________________________ OpenSSL Project http://www.openssl.org User Support Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
