Hello, > I have a SSL server that uses Openssl 9.7e. It works fine initially > but few hundred connections, lots of SSL hanshake > failure occurs. The failure reason given is SSL_ERROR_SYSCALL. After > that I try to retrieve the more error details from the stack using > ERR_error_string_n or ERR_print_errors_fp (sort of wrapper of > ERR_error_string_n) but bott of these give no error information.. There may be many causes and hard to say why this happen without debugging process. First I suggest some "generic" checks: - when SSL_accept() has error print errno value (may be set or no - it is worth a try) - check for memory leaks (from simple "ps -efl" to valgrind for example - check if you have "ERR_remove_state(0)" at end of your thread function (memory leak) - check for file descriptor leaks (with "lsof") - check what syscall gives you error with "strace" (hard way) - check for too small system parameters, some systems may use kernel messages to inform you about that, check this with "dmesg" - check for not properly disconnected tcp circuits with something like "netstat -an | grep -i time | wc -l" - check if you have thread callbacks allocated/initialized Hope this helps.
Best regards, -- Marek Marcola <[EMAIL PROTECTED]> ______________________________________________________________________ OpenSSL Project http://www.openssl.org User Support Mailing List openssl-users@openssl.org Automated List Manager [EMAIL PROTECTED]