Experiencing what I can only call "script lag" on my rental server (yes, it's AI)

[Dorian Gray 10]

I am running DAC on an Art of War Central server, and before I go on, the reason that I am starting a thread in the multiplayer section instead of continuing with my problem(s) in the DAC thread is that it would be rude, and, I recently confirmed with my friend that our DAC mission loads in a few minutes on our little local dedicated servers, compared to the current record of ~20 minutes on my AoWC server. To be clear, I am not "blaming" DAC; there is most definitely some general setting(s) or a combination of settings that I am missing server-side. I have tried many, many things already, such as adding in the bandwidth settings in the config file, and the whole cpuCount, exThreads, and maxmem in the batch file. Through all my tests, my ping was absolutely fine (20-30). One thing I wonder, is if DAC behaves differently on a standalone server to begin with, compared to our local dedicated servers which rely on the entire game install. If it matters, we are actually testing two different missions, on Lingor and Fallujah. So far Lingor is going up to 20 minutes as I said, and Fallujah about 7. If it comes to it, I can link to the missions.

[.kju 3245]

Please post your basic.cfg:

http://community.bistudio.com/wiki/basic.cfg

[sickboy 13]

Perhaps under certain configurations + settings of -exThreads could cause worse script performance, perhaps try a few different -exThreads options.

All DAC initializations/scripts run in scheduled space, and therefore is generally slow, which gets amplified by the other amount of scripts running.

[Dorian Gray 10]

I have already been to that page, and right now my config file doesn't contain those settings because they didn't appear to have any effect.

language="English";
adapter=-1;
3D_Performance=46875;
Resolution_Bpp=32;

I forgot to mention, it's only a 8 player server.

Sickboy, I have tried 7 and 5 already, I will give 3 a go ...

Edited March 10, 2011 by Dorian Gray
added reply to Sickboy.

[Lonestar 11]

http://forums.bistudio.com/showpost.php?p=1870814&postcount=748

You should check your basic.cfg (try default values).

Try a mission without DAC, it may be wrongly set up in your missions.

[Dorian Gray 10]

I'd have to disagree about the DAC setup, since as I said it takes 3 minutes on my machine (which is not meant to host a real server), and 20 on the AoWC server. I'm still interested to know if DAC treats a standalone server differently. My post from the DAC thread that you linked to is precisely what I already tried.

[Lonestar 11]

Try default basic.cfg values: http://community.bistudio.com/wiki/basic.cfg

Your MaxBandwidth value may be too high.

What is the performance with a mission not using DAC?

Try to isolate the possible causes.

Edited March 10, 2011 by Lonestar

[Dorian Gray 10]

Your MaxBandwidth value may be too high.

Once again, I have been to that page already, and I currently do not have any bandwidth settings whatsoever in my cfg, because they weren't making a difference:

language="English";
adapter=-1;
3D_Performance=46875;
Resolution_Bpp=32;

Those lines are generated by default, as the biki says, and I'm sure that the bandwidth settings are optional, as they aren't there in the first place. I have no way of knowing how the server is performing without DAC other than my ping, unless you have other ideas?

[Lonestar 11]

Bandwidth settings optional? I think you're wrong.

Have a look there and download the arma2.cfg sample.

You use your ping to monitor performance on a DS?

Use the #monitor command instead.

http://community.bistudio.com/wiki/Multiplayer_Server_Commands

Of course we are talking about a remote dedicated server, not a local DS? Aren't we?

Edited March 10, 2011 by Lonestar

[Dorian Gray 10]

New York. I'll get back to you shortly regarding the sample and #monitor.

[rexehuk 16]

Just thought id put this in here... as people are talking about bandwidth a lot and they are right.

Over the last few days I've done tests with min error rates and the results are pretty substantial. The variation from 0.01 and 0.0001 error rates was over 32 FPS on serverside FPS performance (0.01 giving the most FPS).

Rest of the settings as follows:

MinBandwidth=10000000;

MaxBandwidth=2147483647;

MaxMsgSend=512;

MaxSizeGuaranteed=1024;

MaxSizeNonguaranteed=128;

If I get time I will be doing a bit more testing... hopefully plotting it on a graph if I can.

[Dorian Gray 10]

I would like to thank everyone for their support in this matter.

I have indeed isolated the problem; it's good and bad news. As I stated in my first post, I was never blaming DAC in the first place, but now I have something to blame, and that is Art of War Central. This is not a "bash", more of a warning - their current system simply does not work "out of the box".

I come from leasing many UT2004 servers, which I understand, are entirely different animals, but they always worked just fine (at least the company I used) after adding custom content (maps, mutators, TC mods). There was never a severe drop in performance after adding content. Over a five year period I had about 10 UT2004 servers, with again, excellent performance, so perhaps that sparked some ignorance in me - assuming that moving on to ArmA 2 would as pleasant an experience ...

However, I have since switched to JestServers, with absolutely no issues running the DAC Lingor mission in question.

To go in order:

Perhaps under certain configurations + settings of -exThreads could cause worse script performance, perhaps try a few different -exThreads options.

All DAC initializations/scripts run in scheduled space, and therefore is generally slow, which gets amplified by the other amount of scripts running.

JestServers actually contains an option in their command line builder to use exThreads, and I set it to 7 right off the bat, before the very first reboot with my custom mod string. Does anyone know if DAC actually needs that? It is common knowledge that an optimal value is chosen automatically. I spent 10 bucks more to get an 8 core processor instead of 4 like I did with AoWC. Whether 8 vs 4 is significantly better for ArmA 2 and/or DAC, I also do not know.

Bandwidth settings optional? I think you're wrong.

Have a look there and download the arma2.cfg sample.

You use your ping to monitor performance on a DS?

Use the #monitor command instead.

http://community.bistudio.com/wiki/Multiplayer_Server_Commands

Of course we are talking about a remote dedicated server, not a local DS? Aren't we?

Thank you very much for the tip on using #monitor. I had seen that wiki page but happened to overlook that section. However, in this case we were both wrong. From the biki:

These options are important for performance tuning

That's why I referred to the settings as optional.

Initially the DAC Lingor mission would devastate the AoWC server down to 10-15 fps, without adding in any bandwidth settings. After I tried several settings, I was still lucky if I could break 30. With JestServers, I did not need to add anything and the server was running at 45 fps on average. DAC finished in under 5 minutes, exactly as intended. This server is also in New York.

From my griping in the DAC thread:

No, the latest is 78413

http://www.arma2.com/beta-patch.php

I had linked AoWC to that page and they said that they could not install it because "it was looking for the whole game, which we do not have". So the most I could do was have them install that hotfix. I almost forgot, JestServers supports the betas.

Once again, thanks for the support, heed my warning, and hopefully my testimonial.

[Jest 10]

Happy things are working out!

Let us know if you have any issues.

[teilx 4]

Dac has problems on Lingor and Fallujah(some more Maps) with creating waypoints, are my expierence.