A faster login process for DayZ

Fred, not sure if this is an issue cause by your changes or not BUT players on my servers when they break thier legs, they fix them with morphine etc. but when they disconnect and join later on.. they have broken legs again.
 
Fred, not sure if this is an issue cause by your changes or not BUT players on my servers when they break thier legs, they fix them with morphine etc. but when they disconnect and join later on.. they have broken legs again.

I confirm this. Lots of complaints.
 
Fred, not sure if this is an issue cause by your changes or not BUT players on my servers when they break thier legs, they fix them with morphine etc. but when they disconnect and join later on.. they have broken legs again.
... i know this symptoms, the reason is simple:
Because relogin is very fast now, it is possible to relogin BEFORE some slow running DB updates are complete.
So you may read older data from DB if you fast relogin.
I changed in publicEH.sqf some related eventhandlers to speed up the DB updates.
If you use this faster EH's already and still have problems, the only tip i can give you, is speeding up your DB!!!!
Especially write access speed is important here, you can do something with your InnoDB settings and/or use ramdisk for logfiles for example.


Let me know your results.
 
Fred. I personally think it's not related to the DB speeds.
I think it's still the arma2oaserver.exe that's causing slow read/write's

Because I run multiply other things from the same database, runs perfect and fast even when the dayz server is full and slow. The website and stuff are still as fast as normal.
Thats just my thoughts though. As MySQL isnt that bad at all.

For example. How longer the server runs, how longer it take for write's:
PHP:
Database.MySQLConnection: [Trace] Execute [33 ms] SQL: update `survivor` s set s.`medical` = '[false,false,false,false,false,false,false,12000,[],[0,0],0,[152.35,153.705]]' , s.`state` = '[\"BAF_ied_v4_muzzle\",\"amovpknlmrunsraswrfldfr\",38]' , s.`Worldspace` = '[77,[7707.21,3243.38,0.002]]' where s.`id` = 22140
33 MS for a simple update query that normally takes 0 ms ?

And with a simple server restart ( Which doesnt restart MySQL ) It will work fine again. For me a good example why it isnt DB related.
 
Fred. I personally think it's not related to the DB speeds.
I think it's still the arma2oaserver.exe that's causing slow read/write's

Because I run multiply other things from the same database, runs perfect and fast even when the dayz server is full and slow. The website and stuff are still as fast as normal.
Thats just my thoughts though. As MySQL isnt that bad at all.

For example. How longer the server runs, how longer it take for write's:
PHP:
Database.MySQLConnection: [Trace] Execute [33 ms] SQL: update `survivor` s set s.`medical` = '[false,false,false,false,false,false,false,12000,[],[0,0],0,[152.35,153.705]]' , s.`state` = '[\"BAF_ied_v4_muzzle\",\"amovpknlmrunsraswrfldfr\",38]' , s.`Worldspace` = '[77,[7707.21,3243.38,0.002]]' where s.`id` = 22140
33 MS for a simple update query that normally takes 0 ms ?

And with a simple server restart ( Which doesnt restart MySQL ) It will work fine again. For me a good example why it isnt DB related.

... i will not start a DB discussion here, but one point you should notice:
Your hiveExt.dll is probably queueing your DB writes, look in your hiveext.log and you will see always [oms] for writes, even if your MySQL updates in real need 33ms ...

So my tip is still, try to optimize your write speed.
But you are right, there could be other things too, which are responsible for delays.
 
I'm also not lookin for a discussion with you. I like your work, you should know that right, I believe what you are saying and posting here. :)

I will try to optimize my DB. and report if anything changes.
 
No, sorry killzone, cant find it in dayz scripts.
It could be a message from the engine it self and seems related to zombie hits?

BTW: ... to find some strings in directorys/files i use notepad++, it has very powerfull search functions ...
 
The difference between a conservative version of the full?
... the "conservative" publicEH.sqf is a "fallback", which you can try if you encounter problems with experimental ...

However, experimental unloads server a lot by avoiding useless code execution.
 
the proof that earlier dayz code dint have any of this

before
Code:
class zZombie_Base : Citizen1 {
        scope = 2;
        glassesEnabled = 0;
        vehicleClass = "Zombie";
        displayName = "Zombie";
        fsmDanger = "";
        fsmFormation = "";
        zombieLoot = "civilian";
        moves = "CfgMovesZombie";
        isMan = 0;
        weapons[] = {};
        magazines[] = {};
        sensitivity = 4;
        sensitivityEar = 2;
        identityTypes[] = {"zombie1", "zombie2"};
 
        class TalkTopics {};
        languages[] = {};
 
        class Eventhandlers {
            init = "_this call zombie_initialize;";
        };

after

Code:
class zZombie_Base : Citizen1 {
        scope = 2;
        glassesEnabled = 0;
        vehicleClass = "Zombie";
        displayName = "Zombie";
        fsmDanger = "";
        fsmFormation = "";
        zombieLoot = "civilian";
        moves = "CfgMovesZombie";
        isMan = 0;
        weapons[] = {};
        magazines[] = {};
        sensitivity = 4;
        sensitivityEar = 2;
        identityTypes[] = {"zombie1", "zombie2"};
 
        class TalkTopics {};
        languages[] = {};
 
        class Eventhandlers {
            init = "_this call zombie_initialize;";
            local = "diag_log ('Locality Event');if(_this select 1) then {[(position (_this select 0)),(_this select 0),true] execFSM '\z\AddOns\dayz_code\system\zombie_agent.fsm'};";
        };

... hey, very good catch ...

My notepad++ didn't find it, because of binarization.
You should contact R4ZoR49, he is currently reworking the zombie stuff.
 
Edit: It seems it is triggered for killing zombies too (in my log files it is very frequently with one player).

I think you are right, the execution of this FSM here looks like heavy wasting of resourses.
We should really investigate this !!!
 
Back
Top