Set up local testing with Amazon GameLift Anywhere - Amazon GameLift
Set up a local Anywhere fleet Update and install your game serverTest game session activityIterate on your game serverTransition your game to Amazon GameLift managed fleets

Set up local testing with Amazon GameLift Anywhere

Note

This topic covers local testing for games that are integrated with the Amazon GameLift server SDK version 5.x. If your game uses server SDK version 4.x or earlier, see Test your integration using Amazon GameLift Local.

Use an Amazon GameLift Anywhere fleet and your own hardware to iteratively build and test your game components in a simulated hosted environment. Set up an Anywhere fleet and register a local device to establish a connection to the Amazon GameLift service. Install your game server build onto the device, start a game server process, and test game functionality as needed. You can update your game server build as often as needed to test each new build iteration.

With an Anywhere fleet, you can test using the AWS CLI or with test scripts. If you've integrated a game client with Amazon GameLift, you can run the client on the same local device or on a different device.

Testing locally with an Anywhere fleet is particularly useful for testing your game server integration with Amazon GameLift. You have full visibility into all hosting activity on the local machine, as well as events and logging data.

Note

Are you using the Amazon GameLift plugin for Unreal Engine or Unity? These tools include guided workflows for setting up local testing with an Anywhere fleet. Follow the documentation for Plugin for Unity: Set up local testing with Amazon GameLift Anywhere or Plugin for Unreal: Set up local testing with Amazon GameLift Anywhere.

Set up a local Anywhere fleet

Follow these steps to create an Anywhere fleet for your local workstation. For detailed instructions using either the AWS CLI or the AWS Management Console for Amazon GameLift, see Create an Amazon GameLift Anywhere fleet.

To create the Anywhere fleet

  1. Create a custom location for your local workstation. (AWS CLI or console). A custom location is simply a label for the compute resource you plan to include in your Anywhere fleet. Custom location names must start with custom-. For example: custom-my_laptop. See Create a custom location.

  2. Create an Anywhere fleet (AWS CLI or console). In this step, create the fleet resource with the custom location for your local workstation. See Create an Anywhere fleet.

    Make a note of the new fleet's ID or ARN value. You'll need this value for the next step.

  3. Register your local workstation as a fleet compute (AWS CLI only). An Anywhere fleet must have at least one compute resource to host your game servers. See Add a compute to the fleet. To add a compute to the fleet, you need the following information:

    • A compute name. Each compute in a fleet must have a unique name.

    • The Anywhere fleet identifier. You can use either the FleetID or FleetArn.

    • The compute's connection info. Specify either an IpAddress or DnsName. This is how Amazon GameLift and game clients will connect to game servers.

    • A custom location in the Anywhere fleet.

    Make a note of the GameLiftServiceSdkEndpoint return value. You'll need this value when you update your game server to run on an Anywhere fleet.

Update and install your game server

This task assumes that you've already integrated a game server build with Amazon GameLift server SDK 5.x. The integration process involves adding code to your game server so that it can interact with the Amazon GameLift service to start and manage game sessions.

For an Anywhere fleet, you need to manually configure certain game server settings. On an Amazon GameLift managed fleet, these settings are configured automatically.

To prepare your game server for an Anywhere fleet

  1. Get an authentication token. Your game server must include an authentication token with every communication with the Amazon GameLift service. Amazon GameLift auth tokens are short-lived and must be regularly refreshed.

    As a best practice, create a script to complete the following tasks:

    • Call the AWS CLI action get-compute-auth-token.

    • Store the returned token value where game server processes can retrieve it, such as in an environment variable on the local compute.

    Install the script with your game server on the compute. Set the script to run before starting the first game server process. While game server processes are active, run the script regularly to maintain a valid auth token. All game server processes on the compute can use the same auth token.

  2. Update your Amazon GameLift game server code. When you integrated your game server code with the Amazon GameLift server SDK, you added a call to the action InitSdk(). When the game server runs on an Anywhere fleet, this call requires additional server parameters. For more information, see Initialize the server process and the Amazon GameLift server SDK 5.x for your development language. The server parameters are:

    • webSocketUrl – Set this parameter to the GameLiftServiceSdkEndpoint value that is returned when you register a compute with the fleet.

    • hostId – Set this parameter to the compute name that you specifiy when you register a compute with the Anywhere fleet.

    • fleetId – Set this parameter to the ID of the Anywhere fleet.

    • authToken – Set this parameter to the token that is returned in response to a request to retrieve an authenticaiton token for a compute.

    • processId – Set this parameter to identify a game server process that's running on the local compute. Each concurrent game server process must have a unique process ID.

    The server parameter values that each game server process uses needs to be specific to the Anywhere fleet compute where the process is running. For details on how to get the appropriate values for a compute, see Add a compute to the fleet. As a best practice, set webSocketUrl, hostId, fleetId, and authToken as environment variables on the local compute. All server processes that run on the compute will use these values.

  3. Install the game server build on the local compute. Include all dependencies needed to run the game server.

  4. Start one or more game server processes running on the local compute. When the game server process calls the server SDK action ProcessReady(), the process is ready to host a game session.

Test game session activity

Test your game server integration by working with game sessions. If you don't have a game client integrated with Amazon GameLift functionality, you can use the AWS CLI to start game sessions. Try the following scenarios:

  • Create a game session. Call create-game-session command (or the CreateGameSession API operation). Specify your Anywhere fleet's ID and custom location. This call returns a unique identifier for the new game session.

  • Check game session status. Call describe-game-sessions command (or the DescribeGameSessions API action). Specify the game session ID. This call returns detailed game session information, including the game session status. Game sessions in Active status are ready for players to connect. To get a list of all game sessions for the fleet, call list-game-sessions command (or the ListGameSessions API action).

  • Connect to the game session. If your game client has the ability to join a game session, use the connection information included in the game session information.

Iterate on your game server

You can use the same Anywhere fleet and compute to test other versions of your game server build.

  1. Clean up your existing GameSession. If the game server process crashes or won't call ProcessEnding(), Amazon GameLift cleans up the GameSession after the game server stops sending health checks.

  2. Generate a new game server build. Make changes to your game server and package an revised build.

  3. Update the game server build on your local compute. Your previous Anywhere fleet is still active and your laptop is still registered as a compute resource in the fleet.

  4. Get an updated authorization token. Call the get-compute-auth-token CLI command and store the token on the local compute.

  5. Start one or more game server processes running on the local compute. When the game server process calls ProcessReady(), it's ready to be used for testing.

Transition your game to Amazon GameLift managed fleets

After you've completed development testing and you're ready to prepare for launch, this is a good time to switch over to Amazon GameLift managed fleets. Use managed fleets to fine-tune and test your game hosting resources. Implement your game session placement solution (queues and matchmakers), select optimum hosting hardware (including Spot fleets) and locations, and choose a strategy for scaling capacity. You might also want to start using AWS CloudFormation to more efficiently manage the life cycles of all your game hosting resources, including fleets, queues, and matchmakers.

You need to make a few minor modifications to transition from a local Anywhere test fleet to an Amazon GameLift managed fleet. You can reuse the same queues and matchmakers. Do the following tasks:

  • Change the game server code call to InitSdk(). Remove the server parameters. For a managed fleet, Amazon GameLift automatically tracks this information.

  • Create an Amazon GameLift build resource. With an Anywhere test fleet, you have to manually deploy your game server build and dependencies to each fleet compute. With a managed fleet, you create and upload your game build package to Amazon GameLift, which automatically deploys it to all fleet computes. See Deploy a custom server build for Amazon GameLift hosting for details on packaging your game build files and creating a build resource with files in an Amazon S3 bucket. Don't include scripts that register a compute and get an authentication token, as Amazon GameLift automatically handles these tasks with managed fleets.

  • Create a managed fleet. Create a fleet using the console or AWS CLI, specifying an EC2 managed fleet. This type of fleet requires additional configuration settings, including specifying the build resource and instance types. You alls need to set up a runtime configuration to manage game server life cycle on each fleet compute. See Create an Amazon GameLift managed EC2 fleet for details on creating a managed fleet.

  • Redirect fleet aliases (optional). If you set up aliases to use with your Anywhere fleets, you can reuse the same aliases for your managed fleets. See Create an Amazon GameLift alias for details on creating or updating an alias.