- Newest
- Most votes
- Most comments
You are right, if you don't have have available processes to host game sessions, your creategamesession calls will fail with that exception. You can use autoscaling, multiple processes per instance etc to get around this problem. Your Fleet metrics should show you all the data you need.
Also ensure you are terminating game sessions correctly to free up the server process for another session.
And you are also right that when running multiple servers per instance, each server needs to run on its own port and accurately report that port via the ProcessParams passed via ProcessReady.
Ok, after some more experimentation and a couple re-reads of other posts, it is my understanding that the allowed amount of Game Sessions is tied directly to the number of Concurrent processes for the game executable configured in the fleet.
The reason my simultaneous create game session requests failed is because there wasn't another available process to run the game.
I also want to say that if you have multiple game sessions, make sure your server picks a different port for each game session, otherwise you'll have some potential cross session communication.
If someone else can weigh in on whether or not this is accurate, that would be awesome.
thanks
Also pay attention to the values for GameSessionActivationTimeoutSeconds
and MaxConcurrentGameSessionActivations
in your CreateFleet settings, both can impact the availability of new sessions.
Relevant content
- Accepted Answerasked 2 months ago
- AWS OFFICIALUpdated 16 days ago
- AWS OFFICIALUpdated 5 months ago
- AWS OFFICIALUpdated 5 months ago