Some days ago (2015-02-04) I've configured all servers to be steam group exclusive:
- Code:
-
sv_steamgroup_exclusive 1
I wanted to see if it makes a difference.
And it seems it did.
It takes some time till the steam master servers replicate that server setting.
So either I would have to turn the servers off for some time so that they get purged from the master server cache, or it seems that it also works by just waiting some time (maybe the root servers will refresh servers settings in some interval)
With this setting no random lobbies will connect to our servers anymore.
Only lobbies started by players from our steam group could still be assigned to our servers or you could enfore the lobby connect to one of our servers as described
[You must be registered and logged in to see this link.] with e.g.
- Code:
-
mm_dedicated_force_servers 81.169.228.209:27016
There are pros and cons about that setting.
Often lobby players are confused if they are assigned to one of our custom servers.
They think they have a private game and kick other players who are joining later.
Also lobby players are confused about custom server features like "buy system" or automatic difficulty settings (more CI and SI etc)
Often I see them connecting via lobby to one of our servers, just to leave a few seconds/minutes later because they expected e.g. a friends-only lobby on any official server. So they create server traffic just for launching a map, connecting to the server, then leaving quite quickly (e.g. very often on GunGame. They complain about "There are no Guns", "What a shit", etc. Then they leave.)
On the other hand some lobby players could find our servers occasionally via lobby and come back later via connect to explore the servers further more because they like the custom settings etc.
So I'm not sure which way to go to attract more long-term custom-liking players.
Maybe we should make some advertisements somewhere^^
Here is a server CPU overview for one week.
You can see that the CPU usage is getting lower after I have enabled that setting mentioned above (at "04" on the x-axis):
[You must be registered and logged in to see this link.]Also the dayly CPU graph is lower and it contains far less spikes because there are no more random lobby connects/disconnects shortly after:
[You must be registered and logged in to see this link.](I don't have a dayly graph to show the spikes before activating this setting)
Well, I just wanted to share that newest result/finding with all of you.
CU soon
Al