Why are the RRSF sessions failing?

The return code and reason code are documented in Unix System Services Messages and Codes manual. In this case, they are indicating that a TCP RESET packet was received on the connection. There can be many causes for this, but the typical reason is that there is a firewall (or multiple firewalls) in between the two systems and one of them has an idle connection timer configured that terminates the connection if there is no traffic for that period. When the next attempt to send a message to the partner system occurs, the firewall will respond with a RESET.

The solution for this case is to set the INTERVAL value on the TCPCONFIG statement in the TCPIP PROFILE to something less than the firewall's idle session timer. You will need to contact that firewall's administrator to get the current setting (this will typically be 30 to 60 minutes). The INTERVAL setting can be changed dynamically in TCPIP, but the RRSF sockets (including the server side listener) will need to be recycled to pick up the new value.

If that is not the explanation, then collecting a packet trace of the failure from both systems would be the next step (see TechNote 1292013 for directions).