Question Creating a "Command Channel" Client to control a Netcode Host

You can either use custom scene management, or pur the command client in the same scene of the other ones but hiding all the "non-relevant bits". Option 2 is probably easier.

Does this help?

 

Hi just asked this internally to the Netcode team, will get back at you as soon as I have a valid answer

 

Hello @VRS3DGuru,
This is possible but there are a few methods you will want to look into using:

• You will need to determine what kind of client synchronization mode works the best for your project
  • Based on your description I would suggest using an "Additive" client synchronization mode.
• On the command client side, you can use:
  • NetworkSceneManager.VerifySceneBeforeLoading to control what scenes are loaded
  • NetworkSceneManager.VerifySceneBeforeUnloading to control what scenes are unloaded
  • You can also disable scene validation warnings using NetworkSceneManager.DisableValidationWarnings
• Most likely you will want to disable ForceSamePrefabs (you can do this via inspector view in the NetworkManager)
  • More on this below

You should should keep in mind that when a client connects and the connection is approved, it will be synchronized with all spawned NetworkObjects. This means the client will need to "know" how to handle the spawning process for those NetworkObjects. If you have in-scene placed NetworkObjects but don't load the scene, then this can cause issues.

Without knowing how your project is setup and all of the specifics (how many scenes, of those scenes how many in-scene placed NetworkObjects, etc), some of the next recommendations you will have to determine if they work for your project. I would highly recommend reading through the object spawning documentation first.

Since in-scene placed NetworkObjects are a bit of a conundrum to handle (for this scenario) you might think about converting any in-scene placed NetworkObjects over to a Dynamic Spawned (non-pooled) in-scene placed NetworkObject (also read a bit about the "Hybrid Approach" example). In-Scene placed NetworkObjects are convenient in that you can easily configure its settings via the inspector view, but for your scenario you would need to make any in-scene placed dynamically spawned NetworkObject use a ScriptableObject to uniquely configure it or for a more advanced approach handle serializing the "unique properties" via NetworkBehaviour.OnSynchronize. The reason behind the "Hybrid Approach" for your in-scene placed NetworkObjects is that once the associated network prefab is spawned it is synchronized like a dynamically spawned NetworkObject (because it really is being dynamically spawned). The advantage to this is that the network prefab being spawned just has to be registered in the prefab list and doesn't actually have to exist in the scenes (that your command client isn't going to load).

The final thing to consider is using a unique NetworkManager configuration for the command client where you might want to create network prefab overrides that point to "place holder replacement" network prefabs that might not have all of the assets required (i.e. could just be a GameObject with NetworkObject).

You will also want to handle object visibility (really this means "Network Visibility") on the server side where you can make certain NetworkObjects "not visible" to the command client so it doesn't receive messages (i.e. you have an override network prefab that doesn't have the NetworkBehaviours with RPCs and/or NetworkVariables on the command client so you don't want to try and send RPCs or NetworkVariable updates for any instance of that network prefab to the command client).

With all of this said, this should get you pointed in the right directions and you can post any questions you have about the above (or issues you run into here).