Configurable idle timeout for Homer TCP connection #2040
Description
Is your feature request related to a problem? Please describe
Hi,
I'm using rtpengine with --homer and --homer-protocol=tcp to send HEP traffic to a local stunnel instance, which then forwards the traffic (over TLS) to a remote Homer server.
From time to time I see the following log line:
[1765850620.856975] ERR: [U62KpxU9D2/bB5blxZxA/1 port 36493]: [core] Write error to Homer at 172.16.1.198:9062: Broken pipe
After some investigation it looks like the remote side (Homer/stunnel) closes the TCP connection when it has been idle for too long. On the rtpengine host, when the remote side closes the connection, netstat shows the socket in CLOSE_WAIT state. The connection stays in CLOSE_WAIT until the next call, and when rtpengine tries to send the next HEP packet over that connection, the write fails with Broken pipe.
My question: is there currently any way to control the idle timeout behavior for the TCP connection from rtpengine to Homer?
If this is not configurable today, would you consider adding such an option or automatic reconnect behavior when the Homer TCP connection has been closed by the peer due to idling?
Environment
- rtpengine version: 12.5
- OS / distro: bullseye