Hey,
I am making some progress on the multithreaded socket server. Thinks are going well. However; every so often pretty much all (or all of them, I'm yet to check - either way, it's a majority) of the connected sockets end up having SSL_read() performed on them, and this function returns 0. I then run SSL_get_error(ssl, [return value of SSL_read) which gives me back SSL_ERROR_SYSCALL I then run ERR_print_errors_fp(stdout) to check the remainder of the errors, but there are none! And errno immediately after calling SSL_read() is zero (I set it to zero prior to calling SSL_read as well). What is going on here? A majority of the sockets all return this at the same time (a read loop which iterates each socket, comes across this error for all of them) usually within 1-2 times per hour, causing the sockets to be closed (because I pick up a return of 0 as an error). What is going wrong here? 80+ connected clients all get disconnected due to this outstanding issue. The other 99% of the time, the system is functioning correctly and communicating fine. Because it says SSL_ERROR_SYSCALL, I do as the manual suggests and check the error queue - there are no other errors. Errno is also zero. What am I supposed to do to handle this error? Any ideas? Thanks, Andrew
