Last night, race night, we again had Mass Disconnects with the trace file recording a TCPBuffer error. 1668.29s NetComm.cpp 1572: System Message(2): "Error (The socket is marked as nonblocking and the requested operation would block.) sending to 27.99.24.91:65227" 1668.29s NetComm.cpp 1572: System Message(2): "Error (The socket is marked as nonblocking and the requested operation would block.) sending to 27.99.24.91:65227" 1668.29s HPWsock.cpp 636: Cannot store specified data in TCPBuffer (not enough buffer space) 1668.29s NetComm.cpp 1572: System Message(2): "Out of outgoing buffer room for 27.99.24.91:65227" With further investigation we found TCP connections maxing out at 500. The server had multiple (hundreds) to 142.250.204.4 which appears to be a google server. When we closed "rfactor2 dedicated.exe" these connections disappeared. Can someone please confirm if this is something irregular or the norm?
Check processes, it looks like you have multiple servers running in background !!! You should see only one rfactor 2 dedicated.exe unless you really start multiple servers ...
Hey @DanRZ , Should we be concerned with the number of connections going to the that 1 IP Address 142.250.204.4? Here is Tcpip\Parameters The server didn't have Tcpip/Interfaces/ID. this is what the server does have.
For the ID that is the big number you see which correspond to the 4 "net" interface : can be wifi, rj45 cards and others ... Check inside all of them. You can probably identify it inside the ID folder ...
I checked on my server and i also don't have TCPNumConnections anywhere ... It is probably set by default and not causing troubles.
Hey Dan, I don't understand what you mean by the "4"net" interface"? I checked each of the items under the Interfaces folder, but I don't see an ID folder? Can you please explain again?
Hey @DanRZ , If you run a server, can you please check your TCP connections and see if you have any pointing to these addresses? We have done a fresh reinstall from Windows Server up, and we have 2 dedi's running (v's 4 previously) and we have half of our TCP connections in use. Please share your TCP connection screen?
I will do it later ... and see if i have this kind of connections. If you look at svchost.exe it may be not linked to the rf2 dedicated ...
I checked with a friend who hosts a dedi using Windows 10 Pro, and they did not have the problem. we are running Windows Server 2019. When we turn rf2 dedi off, the TCP connections drop.
Check all the plugins you have and disable them one by one. LiveRacers, Weather, or other plugins may use some external connections and use some ports.
Thanks @DanRZ , we have done that. Turned them all off and nothing changed. We also asked another contact who runs a server and they have the same thing happening. Their server has multiple connections to 142.250.70.132 which is also Google.
I have no connections on distant port 80 ... I've made a montage of my connections. 2 servers running ... I'm also on Wondows Server 2019.
I rechecked and i have a few on port 80. To a23-217-250-58.deploy.static.akamaitechnologies.com (cloud provider).
Thanks. Interesting. at least we have replicated it, although now it looks like it is not a Windows Server 2019 problem as you don't have it. Both of the servers we have seen the problem are in Australia.
@TheSecret1 This is just a Shot into the very blue. Can it be, that you Broadcast the Races on the Server and use an external Website for it?
We also have the multiple tcp port connections to google on our server (also Windows Server 2019), checking the resource manager showed 500 tcp connections. I then created and outgoing rule in windows defender to block port 80 and now, no more google connections (Google ones were the only port 80 connections) and tcp port connections are down to 20. Only downside I see for us atm is if we wish to use a web browser on the server then that is now not possible....r2la still works fine because default port for it is 8000. EDIT: half hour later can access http and https sites via browser on the server...
