Multiplayer consistently doesn't work for certain pairs of players: Perhaps if they are on the same gateway?

Game Details:

  • v0.9.e
  • Online
  • Server: US WEST (player1) + US WEST (player2)

What I did:

  • Player1 (host) invite Player2 to party
  • Player1 starts a monolith
  • Player1 enters monolith
  • Player2 attemps to portal to player1

What happened:

  • Player2 get an infinite Area Loading notification in the middle of the screen
  • On player1’s screen, player2 is show inside of the monolith
  • Meanwhile player1 completes the monolith and portals to post-monolith reward area
  • Player2 after many minutes alt-f4’s the game

What I expected to happen:

  • Player2 is able to enter the Monolith and play

What is the possible impact of this bug?

  • Player2 game is unplayable while the bug is happening
  • Additionally, Player1 cannot create a new monolith while player2 is stuck in limbo and has to be kicked from the party for player1 to continue

Workarounds?

  • None known

Is this consistently reproducible:

  • Yes and no
  • Yes, but only with specific players (possibly that connect normally to the same server I am on)
  • If I do this same thing with (tested with 2 other players) that normally connect to OTHER servers, the problem doesn’t exist

For what it’s worth, the game client shows either around 20ms or around 40ms to the US WEST servers for me. If it’s necessary, I can dig in to try to determine what aws elb ip(s) i’m connecting to. Just let me know if that is needed.

Thanks!

I don’t know anything about your architecture, but from what little I’ve been able to glean publicly the game uses AWS elastic load balancers to front the game traffic.

I do have experience with AWS + cloud services, so given that I don’t seem to have issues partying with people when they are connected to different gateways, I wonder if there is a (missing) back-end ACL to allow traffic between say, a-side (player1) and b-side (player2) traffic within the same region or at least between same-services across clusters within the same region? Or, a case where backend services are resolving a short service name and trying to talk to a cluster-local service when they really need to be talking to a cross-cluster service (in the same region) + (everything is a dns problem?).

Anyway, I hope this gets sorted soon because I’d very much love to play with more of my friends!

Thanks.

Happening to me as we speak. My buddy is in Canada and I’m in the states. whenever I TP to him I hit the infinite loading screen. As you mentioned, he can see me there and do mono’s without issue while I sit there until I exit.