I’m using a rack mounted PortServer with all ports configured to be used as Real Ports. I’m using 8-pin RJ45 cables (with 3 wires connected only), without any flow control. I use the ports from services running on Windows 2000 systems; the port is open when the service starts and closed when the service stops (which should be never, except for maintenance problems or updates).
At the beginning, everything works fine, but if no command is sent to the real port for some time (let’s say 2 or 3 days), messages cannot be sent anymore through this port (simply nothing happens on the RJ45 side, without errors). Usually, restarting the service allow to take control of the port again.
Furthermore, if a port is not used for a longer period (around 1 month, for example), without any software accessing it, use it becomes impossible without rebooting the port server (again, no errors are raised, simply nothing happens).
Any idea?