Cannot Run Multiple Dedicated Servers in the Same Physical System?

[lobster-san 0]

Hi,

 

I'm having issues hosting more than one dedicated reforger server on the same system, behind the same public IP. I can run a single server without problems, but if I try to launch a 2nd instance, with different configs, that 2nd one will cause the 1st one to shut down. The last log messages left behind by the server getting shutdown are the following, verbatim:

BACKEND (W): DS Room/ Server not exist anymore, shutting down...
BACKEND (E): [RestApi] ID:[16] TYPE:[EBREQ_LOBBY_DSHeartbeatS2S] Error Code:404 - Not Found, apiCode="DedicatedServerNotFound", message"Dedicated server not found"
RPL    : IReplication:: Finishing...
RPL    : rpl::PiP::ProcessNetToGame
 RPL    : IReplication: Finished
RPL    : Pip::Destroy
ENGINE : Game Destroyed.
RPL    : Pip::Destroy

This is not a port binding/conflict/forwarding issue. I can run, join, and play in each instance independently, as long as it's the only one running. (Group members outside the LAN can also join and play just fine.) For example, here's relevant info from each instance's .json config file:

Server 1

{
	"bindAddress": "0.0.0.0",
	"bindPort": 1989,
	"publicAddress": "",
	"publicPort": 1989,
	"a2s": {
		"address": "0.0.0.0",
		"port": 17777
	},
	// ...
}

Server 2

{
	"bindAddress": "0.0.0.0",
	"bindPort": 2001,
	"publicAddress": "",
	"publicPort": 2001,
	"a2s": {
		"address": "0.0.0.0",
		"port": 17778
	},
	// ...
}

I'm speculating, but based on log messages each server has about successfully registering with the backend, using <ip>:<port>, it seems as if the instance that registers itself last with the BI backend causes the 1st one to be removed from the backend, which then causes the 1st server's attempt to find/keep itself in the backend on an interval (i.e. the "heartbeat") to fail and shut down.

 

Regardless of whether my speculative root-cause is correct or not, is there a way around this so I can host both instances simultaneously?

 

I'm just trying to host 2 (password-protected) dedicated servers/instances for my group, on the same physical system, with different mod sets, b/c it would be easier than having to SSH and modify JSON config files every single time before playing (e.g. changing mods, etc). The physical system is in my home, and thus behind the same public IP (which may be relevant); this is not through a 3rd party provider. This cannot be changed.

 

Help is appreciated.

 

PS: If not currently possible, BI taking note of this need and/or requirement would be nice.

[ozzbik 71]

Are you using the same executable to start the server or a copy?

[SirSkunkalot 0]

I have the same problem. I have the servers in c:\servers\Reforger1 and c:\servers\Reforger2 and configs in ..\My Games\ArmaReforger1\config1.json and  ..\My Games\ArmaReforger2\config2.json My ports are 2001 and 2002 a2S port are 17777 and 17778. I see port are correct when the servers start but I get this message on server 1 when I start the second one:

DS Room/ Server not exist anymore, shutting down...

[RestApi] ID:38 TYPE: [EBREQ_Lobby_DSHeartbeatS2S] Error code 404 - not found, apiCode="DedicatedServerNotFound", message="Dedicated server not found"

[lobster-san 0]

On 9/21/2024 at 3:45 PM, ozzbik said:

Are you using the same executable to start the server or a copy?

I'm not sure I understand your question here. I'm using the server executable you download from Steam..

[lobster-san 0]

On 9/23/2024 at 4:00 PM, SirSkunkalot said:

I have the same problem. I have the servers in c:\servers\Reforger1 and c:\servers\Reforger2 and configs in ..\My Games\ArmaReforger1\config1.json and  ..\My Games\ArmaReforger2\config2.json My ports are 2001 and 2002 a2S port are 17777 and 17778. I see port are correct when the servers start but I get this message on server 1 when I start the second one:

DS Room/ Server not exist anymore, shutting down...

[RestApi] ID:38 TYPE: [EBREQ_Lobby_DSHeartbeatS2S] Error code 404 - not found, apiCode="DedicatedServerNotFound", message="Dedicated server not found"

Yeah, I think what I speculated in my OP may actually be correct, or close enough to to it:

 

On 9/21/2024 at 1:24 AM, lobster-san said:

I'm speculating, but based on log messages each server has about successfully registering with the backend, using <ip>:<port>, it seems as if the instance that registers itself last with the BI backend causes the 1st one to be removed from the backend, which then causes the 1st server's attempt to find/keep itself in the backend on an interval (i.e. the "heartbeat") to fail and shut down.

 

[zacekjakub 1]

I can confirm this behavior: hosting multiple servers on the same IP seems to cause conflicts, where starting a new server disrupts the previous one.