Reporting thread for Server Test

[Buzzsaw]

Hello All

Thanks to everyone who participated.

We went through a series of stress tests... with the maximum number of players on the server... and then adding AI aircraft flights till the server crashed.

We will be analyzing the logfiles to determine where the problems were and what can possibly be improved.

Any comments or experiences are welcome here:

 

https://theairtacticalassaultgroup.com/forum/showthread.php?t=30019&p=323754#post323754

 

Thanks to all.  ?

[354thFG_Leifr]

How many did you top for player/AI?

[Buzzsaw]

14 hours ago, Leifr said:

How many did you top for player/AI?

The point was not to get the maximum number of players online... we were using ATAG Server 2 which has lower bandwidth... if we wanted the capacity to have more players online we would have used the higher capacity/bandwidth ATAG Server 1.  As it was, we topped out at around 110 players on ATAG 2.  In previous less structured and less monitored tests with ATAG  Server 1 we have topped out at around 135-150.

 

The purpose of the exercise was to collect detailed logfiles which would allow us to isolate causes and determine what errors in the code might cause server crashes, player disconnects or lag effects like 'rubber banding'.  So with that in mind we used a lower bandwidth/capacity server so the test would encounter any issues sooner and with a requirement for fewer players.  The other factor was determining exactly the effect AI flights had on player experience and server performance/failure.  So we gradually introduced increasing numbers of AI to the server once we had maximum numbers of humans flying... until there was a crash.

 

We are hoping the data generated can allow us to refine the already excellent netcode in CoD... as well as pointing to potential solutions for AI aircraft/object load effects.