Chris Griffin
I searched the forums and didn't find anything so here goes:
VM running Win7 w/ rF1 dedicated server running. This machine handled rF2 dedis without issue. What we are suffering from is random/erratic disconnects in rF1. PC now for testing purposes is connected directly to the internet using a PPPoE "dial-up" connection with firewall disabled (yes we're that desperate to figure this out). This has nothing to do with port forwarding.
Running simultaneous packet captures on client and server which are started one right after another so the timestamps are within a second of each other. Good bi-directional UDP traffic following then the following is seen:
- rF1 dedicated server uptime jumps ahead 50 seconds
- server packet capture timestamps jump ahead 50 seconds
- client packet capture timestamps remain unchanged
- rF1 dedicated server immediately sends TCP FIN packet
- client and server negotiate termination of connection
Reproducible 100% of the time. We are trying to figure out why the server uptime and what appears to be TCP/UDP timers would jump ahead suddenly and somewhat randomly. Win7 was setup to sync with time.windows.com but that was disabled (no reboot since) but it seems that is also a weekly activity by default. This is acting more like NTP is running and the system time is being updated constantly, sometimes having to move ahead 50s. We've been able to connect anywhere from 2min to 20min, but never more. We will always get disconnected at some point.
Looking at what is happening the disconnection makes sense. One second your timeout timers are at 0.0001s (you just got a packet from the client) and then a microsecond later the timers are at 50.0001s. That is going to see the server immediately disconnect the client thinking it has not gotten a packet in 50 seconds, which would be appropriate behavior. We can see however in the packet captures that we have full bi-directional traffic on server and client milliseconds before the disconnection notice is sent to client.
Any ideas?
VM running Win7 w/ rF1 dedicated server running. This machine handled rF2 dedis without issue. What we are suffering from is random/erratic disconnects in rF1. PC now for testing purposes is connected directly to the internet using a PPPoE "dial-up" connection with firewall disabled (yes we're that desperate to figure this out). This has nothing to do with port forwarding.
Running simultaneous packet captures on client and server which are started one right after another so the timestamps are within a second of each other. Good bi-directional UDP traffic following then the following is seen:
- rF1 dedicated server uptime jumps ahead 50 seconds
- server packet capture timestamps jump ahead 50 seconds
- client packet capture timestamps remain unchanged
- rF1 dedicated server immediately sends TCP FIN packet
- client and server negotiate termination of connection
Reproducible 100% of the time. We are trying to figure out why the server uptime and what appears to be TCP/UDP timers would jump ahead suddenly and somewhat randomly. Win7 was setup to sync with time.windows.com but that was disabled (no reboot since) but it seems that is also a weekly activity by default. This is acting more like NTP is running and the system time is being updated constantly, sometimes having to move ahead 50s. We've been able to connect anywhere from 2min to 20min, but never more. We will always get disconnected at some point.
Looking at what is happening the disconnection makes sense. One second your timeout timers are at 0.0001s (you just got a packet from the client) and then a microsecond later the timers are at 50.0001s. That is going to see the server immediately disconnect the client thinking it has not gotten a packet in 50 seconds, which would be appropriate behavior. We can see however in the packet captures that we have full bi-directional traffic on server and client milliseconds before the disconnection notice is sent to client.
Any ideas?