An existing connection was forcibly closed by the remote host

[Viper]

Hi

I would like some assistance with this. I have a distributed application that uses TCP .NET Remoting for communication between the server and client computers.

I am pretty sure the issue is network related because the application is currently running at multiple clients, it is running in instances where the UI on the client connects to the server over a TCP port over the internet. In this client's case, they have two sites in two different cities. The main site (where the server is located), the users connecting to the server are working fine. Even I can connect from my office if I connect via a VPN connection to their network. However, the other users from the other town, get the error "An existing connection was forcibly closed by the remote host" and their IT cannot seem to fix this.

Any ideas on how to fix this?

They've already done the following tests:

• They can ping the server from the other site's computers
• They can get an open connection when attempting a Telnet session
• When using a sniffer to monitor the connection, they can see traffic flowing back and forth from the user machine to the server
• On the firewall they have a rule allowing the entire site's DHCP IP range connection to the entire server IP range at the main site on all ports

 

[InertiaM]

My guess would be firewall-related somewhere along the line, either on the PC or the physical device itself.

Can you try changing the port to something other than one the firewall allows, and see what happens then? Does the error change? Check the firewall log to see what it is trapping. Then change the port back to the correct one, and check the log again.

 

[Viper]

I've asked them and they tried using a different port and they tried connecting via Singleton and SingleCall mode and still the same error. Apparently ALL the ports are open between the computers at the site and the servers at the main site.

 

[Viper]

For anyone else who has this issue. They did a full investigation from their side and eventually found the problem. In this instance, they said the following:
"We did find that the network configuration on the MDD side in Durban was not correct. The MDD distribution switch need to have the routing enabled from the user IP range to the server in order to establish a successful connection. This was added and the user could connect."

I have no idea what that means, but if you give this to the IT / Network people, they could perhaps figure out how to solve.