Gamelift Managed Linux Servers Lock Up (Question includes a workaround)

Question

There appears to be a bug in GameLift Managed Linux hosting where standard output is redirected by the process that starts the server executable and is not being read from which causes the server to hang once the buffer fills up.  This is happening with a normal UE4 server.

A workaround is to not use standard output or redirect standard output to /dev/null in the game server executable.  I don't know if this is just an issue with my server or if other people are having the same issue which is why I am asking this question and providing a workaround.

If anyone from AWS can take a look at the server wrapper and see if this is an issue with all managed servers it would be helpful.  It should be a fairly easy fix to implement in the wrapper code.  Hopefully this helps anyone having the same issue.

Answer

Hello,

Thank you for reaching out to us regarding the above query. As mentioned by you, For a [long term solution](https://docs.aws.amazon.com/gamelift/latest/developerguide/logging-server-messages-custom.html), the recommendation is to redirect stdout to a file, and provide the log path to GameLift via LogParameters. This way you can access the log files via the API [GetGameSessionLogUrl](https://docs.aws.amazon.com/gamelift/latest/apireference/API_GetGameSessionLogUrl.html) without having to ssh onto the host. You can periodically clear the buffer using fflush which moves the stream to the desired destination (console in case of stdout / file in case stdout has been redirected to a file).

Having said that, In order to troubleshoot the behaviour being observed, please feel free to open a support case with AWS using the following [link](https://support.console.aws.amazon.com/support/home#/case/create) as we would require specification of your setup.

Thank you

Gamelift Managed Linux Servers Lock Up (Question includes a workaround)

相關內容