"Server query timed out" in direct connect only sometimes

Hello! We’ve just switched server hosts from Hostwinds to OVH for an update. Once the switch was completed, we noticed that some users were unable to direct connect to our servers. This seems to happen to everyone, in different frequencies (maybe in proximity to the actual host?), but not all the time. Even if this happens, people are still able to connect using the connect command.

Our Server Information

  • OVH Oregon Datacenter
  • Windows Server 2019
  • Specs
    • Dedicated Server
    • AMD Ryzen 7 3800X
    • 64GB DDR4 @ 2666MHz
    • 1 Gigabit up/down
  • FXServer version 2744

What we’ve tried

  • Making specific exceptions for FiveM ports outbound (even though windows defender default allows it)
  • Making brand new server licenses
  • Switching interfaces on endpoint_add_tcp and endpoint_add_udp to just our main IP instead of 0.0.0.0
  • Removing the server icon to decrease payload size on info.json

With all of this done, we still couldn’t figure out what the problem was. People were perfectly able to direct connect to the previous host, so we’re wondering if we’ve setup something improperly or if it’s a problem on OVH’s side.

Additional Information

  • Some users, after disconnecting from our servers, see the server token in their direct connect text box (the same one at cfx.re/join/{token})
  • Sometimes while trying to GET one of the *.users.cfx.re urls, it returns with a 504. Not all the time though.
  • Although we don’t have enough information yet to say for sure, but it seems there is a correlation between your distance from Oregon and the amount this happens.
  • Our users are still able to see our servers in the server list, even if they can’t direct connect. If they click in the server list they are able to connect.
  • Viewing the server list then going back to history to direct connect usually temporarily solves the problem.

Alright, after more investigating we figured out it was an OVH firewall only on game servers. This firewall was stopping some, but not all, UDP packets from getting to our server. I figured that if it was a firewall issue, then it’d happen all the time but apparently their firewall isn’t that good lol.

If anyone else is experiencing this, I’d say make sure to check your firewall (if any) on your local server and for the host to see if they’re blocking UDP packets.

2 Likes

Have you reconfigured your OVH firewall?