maciejs
-
Posts
213 -
Joined
-
Last visited
Posts posted by maciejs
-
-
Hmm, you should not preprocess at all (the only exceptions are emblems, but dedicated server should not load these). I'll check it here and make sure it works as expected.
-
Can you elaborate on "this issue" (WAR number ideally, if it's a crash)? The original crash has been fixed a while back.
-
On 3/17/2018 at 2:10 AM, NoSpax said:
Either those are servers ran by the Launcher on a bad connection or they set them up in a way, it conflicts.
On my side, i never connect to a dedicated in my Europe region (even my own dedicated to test) , my game turns into a p2p host, as I am missing the "leave conclave" button when a round ends. Clients can leave. I can only leave in the lobby screen, where i see both teams before a match. I think this is how dedicated server work, as they never leave the conclave.
That's interesting, if you can send me a server log _and_ your game log when it fails to connect to it, I might be able to spot a problem.
-
They're both 'correct', but yes, the latter one is a little bit more elegant.
-
Yes, you inherit properties from your immediate parent.
-
Yes that looks good, I think. You can try using the "DS config file" method (see the first post) for running multiple instances via Launcher.
-
It's hard to say as it's not obvious from the log itself, I need to open a crash dump. Crashes reported before were something else though and should be fixed.
-
Do you have TeamViewer running by any chance? I see it in the callstack of the crashing thread.
-
We did try a few things, give it a try on the latest build.
-
We're looking at it, but it seems like a driver-level crash at a first glance.
-
Thanks. Crashes at shutdown are a little bit lower priority and this one doesn't seem related to the old one, but I'll take a closer look.
-
Send me your EE.log
-
Hey guys, so I *think* we "fixed" it. I put fixed in quotation marks as to be completely honest, still not sure what the issue was exactly. I spent past 2 days staring at various crash dumps and they didn't do too much sense to me. Seemed like some kind of a very violent stack corruption, but even then it wasn't very logical (values in memory/registers didn't 'fit'). We went through all the changes between the last 2 builds and could not find a culprit either. My only theory was the linker/compiler generated some subtle error somewhere and as much as I don't like it because it seems handwavy, I was out of rational explanations and in the words of Sherlock Holmes "Once you eliminate the impossible, whatever remains, no matter how improbable, must be the truth". That particular crash seemed to be limited to Ryzen CPUs, so whatever the faulty code was doing, it didn't seem to matter for Intels. In any case, as of 2018.02.21.13.07, this particular crash should be gone. If you're still having issues, please post a WAR number here.
-
You sir, are correct. I'll update the guide for now and will look into fixing it properly.
[Edit]...and as of the POE update it should be fixed.
-
Oh, Alt+F4/clicking the X icon shuts down gracefully (it doesn't wait for the game to finish, but it does save), I meant an actual crash, can't explain why it wouldn't save otherwise.
-
I went looking through the code to make sure I remembered correctly, but we do not do it (relax region requirements). I think we did consider it at some point, but Warframe is so popular right now that every region should have plenty of sessions active. If you feel you're being matched with players outside your region, feel free to send me a log from your session, I'll try to investigate.
-
Hmm, no, it's not intentional. Steam or not, the first time you run a dedicated server, it just creates a default LotusDedicatedServerSettings object. Went through the code 1 more time, but I don't see any path leading to that outcome, unless the dedicated server process didn't shut down cleanly and/or was not properly initialized (ie. running 'by hand', not via launcher, with invalid arguments).
-
Good point, I'll change the naming.
-
Yeah, it wasn't technically a networking issues. We've added a bunch of new shader permutations and on some drivers/devices, it would not upload them all in time so when it had to render using one of these shader, we'd stall. If you stall for long enough, host disconnects you because he thinks you're gone.
-
Not to jinx it, but I think we have a fix coming soon.
-
18 hours ago, LordMace2001 said:
Really have to wonder if they will ever fix this
Well the problem that you're having is different than other people here, they can join session fine, but game freezes for too long causing the host to drop them. If you send me your EE.log file I'll try to see what's causing your problems.
-
Yes, DX9 will look worse (less fancy features). You can also try DX10/11 + windowed, it seems to help for some. Are you seeing game freezes that correspond to disconnections?
-
20 minutes ago, OscargamersB said:
I already sent it
Yep, got it, thx. I'll add some extra diag in the next build and will contact you again in order to gather some more information. Appreciate your help.
-
So far I've only received one log that shed some light, but could use some more data points.
Dedicated Conclave Servers: Part II, Advanced Guide.
in Developer Workshops
Posted
It should only consider sessions hosted by your clan, yes. Send me your logs (both server & client), I'll take a look.