Continuing Server Issues (13th Sept)

[PeterReaper]

In another thread talking about these lag issues someone else mentions that the problems stem from the UE3 engine itself, which has supposedly been discussed for a long time now. Have a look for yourself.

These severe lag issues didn’t exist in anything even coming close to this severity before the 1.0 release, and I seriously doubt there are THAT many new players - hell, ancient games like DOOM and QUAKE could handle probably an exponentially larger amount of players - without lag. You sound like a nice person, but also like a fanboy. Please don’t try this lame game of smoke and mirrors on me. The massive lag issue is new! SD obviously introduced something with 1.0 that they need to back out ASAP.

[K1X455]

Were suspecting some server side processing is preventing the transmission of the game stream to client (not sure what it is cause we don’t have the knobs and levers of this game) and the client seems to be having game processing lag spikes as identified here.

We’re suspecting it’s EAC contesting the single core for processing that’s causing these spikes.

[boerhae]

Sorry if this is unhelpful, but I’ve been using Alt+Esc for as long as I can remember, so if you ever need to tab out that’s what I’d recommend. I’ve never encountered any issues with EAC and tabbing out with that so hopefully that’s a temporary fix since even the simple things seem to have broken post 1.0.

[K1X455]

I’ll give your Alt-Esc a try. I haven’t done that so in the interest of fair-go… I’ll reply to this when I get the intended results.

EDIT:

Nope… didn’t made a difference.

Same schitt.

[Khuzen]

What works for me is to just have DB open in one desktop and then switch straight to the other one do other stuff. Getting back into the game takes a bit but I think that’s to be expected. I’ve never encountered any EAC problems using this method.

[K1X455]

How do you switch to the other? Alt-Tab? That’s one of the problems I complained at right here.

[Khuzen]

Oops, sorry my bad. Didn’t know that Alt-Tab were the keys for switching desktops. Never had to use it that way because I’m on a laptop. I just use the gestures on my touchpad (4-finger swipe).

Edit: There is a way to switch desktops immediately without Alt-Tab
As long as you have more than one desktop already set up, you can switch between them using:
Ctrl+Win+Left/Right arrow key
It is basically the same as my left/right 4 finger swipe gesture so I think you should try it, it might just work considering how fussy Windows is nowadays.
i.e. ‘desktop 2’ is to the right of ‘desktop 1’, so you use Ctrl+Win+Right_Arrow to switch from desktop 1 -> 2 and Ctrl+Win+Left_Arrow for switching from desktop 2 -> 1

[Rizkifauzan12]

Compensation please, playing almost for 3 weeks for this laggy game, no dev sign so far

[Khuzen]

These severe lag issues didn’t exist in anything even coming close to this severity before the 1.0 release

Please don’t try this lame game of smoke and mirrors on me. The massive lag issue is new!

Huh, I didn’t know that.

During the beta we have never experienced lag as often as it appears now (all the time). What I am trying to say is that the lag issues we all have been experiencing is not due to a problem before the 1.0 update was released, at least not one that would have shown itself or affect the game during that time.

the server issues only surfaced AFTER the 1.0 release.

Okay, moving on.

The developers most likely would not have been able to simulate the real world stress on the servers post update.

I didn’t say anything about the influx of new players (or lack thereof) being the cause of the lag. I meant that the developers would not have been able to replicate a real world test of the update’s performance as well as accounting for all the other variables and therefore would not have known that these issues would have reared their ugly heads. Although I could have explained/included that by this I also meant that whatever changes were made to Dirty Bomb for the 1.0 update would probably have contributed or quite possibly have been the cause of the lag issues. Of which yes, I agree that the lag we have now is the worst it’s ever been (at least in my experience), due to it’s relentless, tormenting nature.

And what about those times during the beta where players including myself have suffered from those 900+ ping spikes, which usually had us kicked from the match? Lag has always been a concern that has constantly surfaced in the Dev videos. I know that this is also dependent on our own internet connection and that nearly all online games are vulnerable to it, I’m just saying that its a possibility that the 1.0 update was not the sole/root cause of all the lag.

It’s not as if I completely disagree with you, my interests align with yours, bottom line I want this issue to be addressed. We all do. I’m simply giving feedback and stating my opinion in a different way, looking at other possibilities, staying positive and giving encouragement. On top of that, I like to provide justifications/explanations and prior research is a must. If it makes me look like a “fanboy” then so be it. I merely dislike pointing fingers. I am not in full agreement with what you are saying, and I’m not here to repeat what you have already said, unless of course you want a pet parrot to keep you company.

[PeterReaper]

The lag issue seems to be fixed - at least in Europe. No word from the devs on what was causing it or what was fixed…

MAGA

[K1X455]

Nope. Absolutely not.

Even with 8ms ping to local server, game processing lag is unchanged.

[Florisjuh]

Until now I’ve found one server in Singapore which got stuck with the “Deploying in…” button and then went back to the main screen.

[Khuzen]

I was in game recently, ended up joining an Australian server with 200 ping. And it was actually playable. Especially compared to what I’ve been experiencing in servers like Singapore and HK. After a couple of matches I left to join a Singapore server to check. But there was still way more lag in the Singapore servers (even at 40-50 ping) and the severity of the lag for me is considered unplayable; constant rubber-banding and teleportation voodoo magic.

[San_Pedro]

Problem seems to be back for me as well.

[K1X455]

[K1X455]

what kind of error is this?

[OCB-Wildcard]

what kind of error is this?

I do believe you’ve found the brother of the white-backdrop scoreboard plate. You still can’t see anything but Won, Lost, and Draw–just like with its predecessor. However, on the flip-side, it doesn’t hurt your eyes to view anymore–so it’s an improvement at least in that regard.

All jokes aside, it probably is a similar bug to that one; no clue what causes this one though.

[B_Montiel]

Netcode needs to be seriously reworked .

It’s not about netcode. UE3 can be a very potent engine/netcode system as long as you know what to do and maintain things thoroughly. Both of which Multiplay is an absolute joke at.

[K1X455]

UE3 can be a very potent engine/netcode system as long as you know what to do and maintain things thoroughly.

You should understand how UE3 handles replication in multiplayer games. This replication phenomenon works on top of UE3’s netcode and if it’s not set properly; or is integrated poorly by way of porting to other applications (ie, EAC), then it will have a bad impact on how the game performs.

This is an example of how poorly netcode is performing in DB 1.0

From the video, you’ll see in successive sequence the following faults as a result of netcode processing failure:

• 100% hits not registering to the server
• Reload animation completing on client, but not acknowledged by the server and hence the client gets 0 ammunition (multiple times)
• Over 20% of hits not registering to the server

[B_Montiel]

I do fully understand that. But as far as my personal experience goes, I’ve seen those happening in every single UE3 games I’ve played extensively : More than 500 hours on Tribes Ascend and over 1200 on chivalry mediaval warfare where I managed one of the most crowded west-European community servers for more than 2 years. And those two games were subjects of the exact same symptoms when servers were going nuts, with totally different third party systems gravitating around the UE3 engine.

At the start of chivalry mediaval warfare, “official” community servers (which could be set to save global XP in the game), were only available from multiplay. We could have servers binaries for non XP saving servers in the meantime. At some point, most of the active communities came to the point that multiplay provided utter shit quality of service. We’ve ran all the same garbage @ASD has been put through 5 years ago by multiplay, with no tangible results coming from them. Some of the communities started to rent from random server hosting service with far better results. The massive difference was that they were renting an actual hardware server, not just a virtual machine, we could actually restart properly and set the best way we could. At that point Thorn Banner broke the deal off and we finally could rent official servers outside of multiplay.

UE3 netcode does not like high pingers and packetloss. It overloads server memory (one of my clan mate held a server with his own fibre cable broadband, at some point, he got most his RAM usage filled with the UE3 binary for No reason, and forced him to restart the computer). On top of that, the tickrate setting has to be met by the server, even fully crowded, otherwise it will have more or less the same consequences. None of that is currently met by multiplay servers, it never did. It’s not new, 8vs8 always had shittier hitreg than 5vs5 ones. On top of that, we have actually no hands on servers parameters, which could potentially give a handy help on those issues. Yes, it has consequences : you have to choose your playing conditions and limit your attendance : in chivalry, a nicely set server was literally unplayable for somebody with more than 100 ms ping, thus limiting a France hosted server up to Poland and Scandinavian countries.