[RGG] 100 Player PvP Server Test (29/06/2013)

[dale 11]

(Time Zone Helper)

Hey all,

I would like to get some idea of how far I can push our server, and also I would like to get a grip with the configs in regards to optimization, the only real way to do this is by asking if a bunch of you would be generous enough to jump on the server and participate in some organised testing.

I feel that now we are in the beta we can finally have a good go at trying out large servers, I am in no way stating that Arma 3 is now stable but I feel that I can start testing without overdue concern that your going to have massive issues, at least to the point that Arma 3 is concerned.

If you could spare me some of your precious time, I will ensure that you have a mission to play with, and I will get it to a good standard with all the latest vehicles etc and see if we can have some all out war on stratis for a few hours.

It will be PvP orientated, any info we learn I will post up so other server admins can learn what I've learnt, and if nothing turns up then so be it but im sure I can provide some information.

If we get enough sign up ill get 2 servers up.

To sign up please check out here: http://www.12thmechanised.com/forums/viewtopic.php?f=52&t=1337&p=4904#p4904

[ric 1]

A good place to start would be from the bottom up...create a mission with NO vehicles or AI just a PvP and set the max to 100 and see how the server and game play handles.

[dale 11]

Still if anyone has anything they want me to do just say it and ill try my best :), the idea here is to push the server to see what kind of performance we get out of it, there will be much mixing and mashing of buttons.

Hopefully because my server is in the middle ground when it comes to performance we can get an idea on what will be best the best setup.

[ZoneTrooper 10]

Dont know if you allready have expirience with alot of players,but it will reduce the fun of gameplay on a extrem lvl.Today i and the Magnon team raised the player limit to 65 Players on wasteland,and it was just horrible.You cant run 1 meter without being killed its just no fun with playing with so manny players,everywhere are players.And if you wanna run a Domination server 50vs50 teams would defenetly kill the server,just expect gunshots everywhere everywhere are players on a 20 square kilometers map.Right now it makes no sense to raise the player limit above 40-50 players,everything above it cuts the fun in the game,but you can try out 100 players and you will see what i mean.If you run a wasteland server with 100 people,i suspect that the server would totally overload,depending of wich version of wasteland you gonna use,if you choose joschaaps version with over 100 vehicles money,food,base-build/system,it will kill the FPS on the server for every player.

[ck-claw 1]

If you run a wasteland server with 100 people,i suspect that the server would totally overload,depending of wich version of wasteland you gonna use,if you choose joschaaps version with over 100 vehicles money,food,base-build/system,it will kill the FPS on the server for every player.

Exactly!

I'm sure Dale is testing pure PvP = hardly any scripts etc etc

[clifdenhill 1]

Hey can you also try 64 vs 64 maybe? I would participate but I haven't got arma 3 yet :(

[St. Jimmy 272]

Will the mission(s) be AAS, A&D, C&H or something similar? Haven't played good PvP since PR died so I'm curious.

[caico1983 10]

Interesting idea! I intend to participate with some buddies from Community of Tatical Combat Brasil.

I would like to know if the server will have password. And if it will be mandatory to register on your forum to participate.

At last, thanks for the link to the Time Zone Helper.

[ratszo 17]

Have no fear, ric - Dale knows what he is doing.

Yes, Dale and the RGG have run 50 vs 50 in Arma2 for years.

Trusted people to play with.

For sign up, i'd rather just meet on your TS server before the match?

Hope all TvT players take note of this and try to attend.

Good Luck!

[Makar 10]

interesting, I'll try to come in and play...

[dale 11]

The first series of tests will be purely on infantry without scripts, just respawn, for this test I will be restarting the server several times with preset configs each time, and hopefully I will find some interesting info in regards to what direction to take when it comes to the configs.

The second series of tests will be with infantry and vehicles, the idea here again being that I find out how vehicles affect the server, each restart will have a different type of vehicle i.e. Chopper, APC, Truck.

The Third series of tests will be about scripts for example vehicle respawn, Capture locations, basically I hope to get minimal scripts with low impact on the server but have a reoccurring role, thus getting rid of the probability of bad scripts but allowing for an understanding on what happens with scripts and up/down limits.

The final test will be a time test in which I will put up a config which seems to suit the best from all tests, and see what happens to the server over say the period of 1 or 2 hours, this will be the more fun bit to be honest and will provide info on what happens over the course of the server life.

Hopefully we will large amounts of information to dissect and work on to provide a better config for our server.

What we will not cover with in this are: AI, more than 2 teams, copious amounts of scripts, the effects mods have on servers.

The idea is to provide information on the base parts of the game, and how to make them run well on the server and reach as high of a player count as we can, in doing this test you are helping us to understand what affects the server the most and how to limit it.

I will publish my findings, but this isn't going to be a dead cert on how to get the best out of your server, because even after the test I will have to refine it more and more until I find at what point the server is comfortable at 100 players.

And my aim here is to provide still a quality level of gaming, so im not just going to plonk everything to 0 I want to find that optimum point at which the server gives its best but doesn't lose its efficiency and that will be a long struggle, for example even having the players view distance set slightly too high will increase the upload speed by about 1 to 2 kbps which might not look like much, but as the game clutters with dead bodies and vehicles that slowly rises but the answer isn't to just limit you to 800 view distance but there is an optimum distance at which it negates the upload speed and still gives you that feeling of seeing into the distance.

There is a lot to look at and I've got a few days left to do some more research by myself, as I say this test will be the start of figuring out how the server reacts to certain configs and at what points do or do not affect the server.

In regards to sign up, consider it as an agreement that you understand what your going to go through, but no you don't have to sign up just come on TS before hand.

The reason I would ask you to sign up is purely for the fact of presence, if there are fifty people signed up then it seems plausible that we will make 100 sign ups, if its just 5 and the rest are hidden then it seems less likely we will hit 100.

It only takes about a minute to sign up and will help others recognise the importance of our test.

Edited June 27, 2013 by Dale

[Druss29 10]

I agree with Ratso. Numbers are not an issue for gameplay.At both Charlie Foxtrot and now Arma-Wargames , we have had no issue with large numbers playing in an organised Team vs Team. Game play is fine. I remember we were told you cant get over 100 players without a poor performance, In the early days it needed tweaked, But in the end it worked fine.

If your looking for players to stress test, advertise on some of the Clan sites to.

[dale 11]

Okay the the event has started, come join in if you want, no addons required.

[SavageCDN 231]

I'd happily join if you do this again... could not make it last weekend. Also would love to hear your results/observations with these tests.