We found a scenario where a downstream app does not like the way that Camel-Mllp disconnects. After digging into the issue, we looked at the camel code and it appeared that the SO Linger gets hardcoded to 0 ms, and appears to be confirmed by sniffer results we are seeing. In short at a TCP Layer we see the client (us) send a FIN and the server (downstream app) sends an ACK to the FIN, but before they can send us the server FIN, we send a RST. That reset crashes their application and requires manual intervention.