I have been using telnet to get dx spots in RUMlogNG for a long time. Often, the telnet session is lost but immediately logs back in automatically. Sometimes it loses connection every minute or two.
Is there a way to prevent this from happening?
Thank you.
I can reproduce this issue with RUMlogNG on my Mac. Analyzing the network packets with tcpdump, I found that RUMlogNG actively closes the connection and then immediately reconnects if it has not received any spots for exactly 210 seconds. Given that HamAlert pre-filters the spots (unlike regular DX clusters), it’s quite common to not have any spots for more than those 3½ minutes. Also, the TCP keepalives that I implemented to appease firewalls and NAT engines that like to terminate idle connections (see Sometimes not receving alerts for a DXCC - #5 by HB9DQM) do nothing to keep RUMlogNG from disconnecting – it wants to see application layer data.
Perhaps the author of RUMlogNG can make this hard-coded timeout configurable (or add an option for it to be disabled).
I am experiencing the same issue in N1MM.
Maybe HamAlert can send notification (message) every 90 seconds or so? Like “No new spots, this message is to keep your software happy”?
Application layer keep-alive has now been implemented. If no spot was sent for four minutes, the cluster prompt will be resent by the Telnet server.