Game Crash 2.7

[casserole]

After Windows 10 update of today 11th may, the crashes continue unchanged.

 

dcs.log-20210511-175445.zip

[ShuRugal]

two more crashes this evening.  Both flying the F/A-18C in multiplayer.  One a bit after a firing a string of HARMS, possibly at the same time a group of T-90s spawned in.  Second crash occurred during handover to tower on the way back to the carrier.

Logs.zip DxDiag.txt

[Flappie]

Hi @ShuRugal. Again, your DCS is given a very small amount of pagefile and I have no idea why. You really should try setting your Windows pagefile to 32 GB (initial : 32768 / max : 32768). I'm saying this because of the your "function-name not available" message in the crash log. It can be a symptom of a memory issue.

 

I know you don't want to mess with the pagefile. Here's a deal: mess with it now, and if your crashes go away, buy more RAM then remove custom pagefile settings.

 

2021-05-13 23:23:49.315 INFO    DCS: CPU cores: 4, threads: 8, System RAM: 32724 MB, Pagefile: 800 MB

 

2021-05-14 00:42:34.277 WARNING LOG: 1 duplicate message(s) skipped.
2021-05-14 00:42:34.277 WARNING Mission: register_unit(1001057,18554113) : db.units['1001057'] table is missing
2021-05-14 00:42:43.984 INFO    EDCORE: try to write dump information
2021-05-14 00:42:43.985 INFO    EDCORE: # -------------- 20210514-004244 --------------
2021-05-14 00:42:43.986 INFO    EDCORE: DCS/2.7.0.5659 (x86_64; Windows NT 10.0.19041)
2021-05-14 00:42:43.986 INFO    EDCORE: D:\Games\Sim\DCS World OpenBeta\bin\edterrain4.dll
2021-05-14 00:42:43.987 INFO    EDCORE: # C0000005 ACCESS_VIOLATION at 4E92D923 00:00000000
2021-05-14 00:42:43.988 INFO    EDCORE: SymInit: Symbol-SearchPath: '.;D:\Games\Sim\DCS World OpenBeta;D:\Games\Sim\DCS World OpenBeta\bin;C:\WINDOWS;C:\WINDOWS\system32;SRV*C:\websymbols*https://msdl.microsoft.com/download/symbols;', symOptions: 528, UserName: 'ShuRugal'
2021-05-14 00:42:43.989 INFO    EDCORE: OS-Version: 10.0.19041 () 0x100-0x1
2021-05-14 00:42:44.282 INFO    EDCORE: 0x00007FFC4E92D923 ((module-name not available)): (function-name not available) + 0x0
2021-05-14 00:42:44.287 INFO    EDCORE: MiniDumpWriteDump failed. Error: 2147943855 

 

[CarbonFox]

Still getting crashes in SP when I switch between different missions and aircraft. Happened twice when I was trying to fly a Mirage mission after flying the Hornet.

[HotTom]

The constant crashes have made DCS unplayable. Unless you get it fixed, everything else ED is working on is pointless. Don't you understand that?

[ShuRugal]

10 hours ago, Flappie said:

Again, your DCS is given a very small amount of pagefile and I have no idea why. You really should try setting your Windows pagefile to 32 GB (initial : 32768 / max : 32768). I'm saying this because of the your "function-name not available" message in the crash log. It can be a symptom of a memory issue.

 

I know you don't want to mess with the pagefile. Here's a deal: mess with it now, and if your crashes go away, buy more RAM then remove custom pagefile settings.

 

I have recently gone from 16 to 32 GB of RAM.  you can see from my procmon capture that while DCS had requested commit for 98% of my total ram, it was only using 16GB at the time of the crash. There are also virtual no hard faults ocurring, the game is not even requesting swap access, this cannot be a swap issue.

I have the swap file completely disabled on all my hard drives except for the drive I have windows OS loaded on, and the swap size there is the minimum to allow windows to save crash data in the event of a BSoD.

If DCS is having memory issues when i have 16GB of RAM unused, that's a problem with DCS, not my system.

[Flappie]

@ShuRugalI was talking about pagefile DCS needs, not general pagefile needs. Windows can live with no pagefile at all, while DCS cannot (I've seen this before). If you don't want to try what I propose, I leave you to it, good luck.

[ShuRugal]

37 minutes ago, Flappie said:

@ShuRugalI was talking about pagefile DCS needs, not general pagefile needs. Windows can live with no pagefile at all, while DCS cannot (I've seen this before). If you don't want to try what I propose, I leave you to it, good luck.

which makes this a problem with DCS.  If DCS is coded to be intolerant of no swap file, that's bad code.

[Flappie]

I'm here to help you solve your DCS problem. I ask you to try something. You don't want to try. What can I do? I'm out of here.

[XPACT]

2 hours ago, ShuRugal said:

which makes this a problem with DCS.  If DCS is coded to be intolerant of no swap file, that's bad code.

 

That is only partially true, game could manage ram usage better but... truth is even if you have 32GB of RAM or more swap file is there to be your fail safe, without it system would be at risk of crashing and corruption at moment there is no space left in main memory, not to mention many apps commit memory, so basically imagine your programs are using 26GB of RAM that means you have 6GB left at your disposal right, but then you open an app that commits to 8GB and will probably use only 3GB while running, guess what you will run into virtual memory low error because you are 2GB over your commit charge, now you can try and argue that program is written badly, maybe it is but even if you had swap file in this case nothing would be written to it unless you go over real usage of 32GB or close to it but you wouldn't experience any virtual memory low related error if you had swap file. Yes, to contrary belief if you have swap file OS won't just write gibberish to it when it doesn't need to, so drive trashing isn't happening.

 

DCS.exe is currently using only 11GB of main memory but is also committed to another 11GB, basically saying hey I might need that memory and I need to have guarantee that it will be ready if needed, program is basically guarding itself from possible memory run out.

 

 

I also have 32GB of RAM but guess what my commit charge is 34GB with DCS open, so without swap file I would be no go.

 

Edited May 14, 2021 by XPACT

[streaker]

I dont understand this, the crashes happen in menu, only once in game. In menu I get crashes on second flight, always, no matter what plane. I have 5 airplanes and all maps, makes no difference.

[Blinde]

The game is just broken. These workarounds are merely just some fixes which are quite poor excuses for bad coding.

[ShuRugal]

23 hours ago, XPACT said:

 

That is only partially true, game could manage ram usage better but... truth is even if you have 32GB of RAM or more swap file is there to be your fail safe, without it system would be at risk of crashing and corruption at moment there is no space left in main memory, not to mention many apps commit memory, so basically imagine your programs are using 26GB of RAM that means you have 6GB left at your disposal right, but then you open an app that commits to 8GB and will probably use only 3GB while running, guess what you will run into virtual memory low error because you are 2GB over your commit charge, now you can try and argue that program is written badly, maybe it is but even if you had swap file in this case nothing would be written to it unless you go over real usage of 32GB or close to it but you wouldn't experience any virtual memory low related error if you had swap file. Yes, to contrary belief if you have swap file OS won't just write gibberish to it when it doesn't need to, so drive trashing isn't happening.

 

DCS.exe is currently using only 11GB of main memory but is also committed to another 11GB, basically saying hey I might need that memory and I need to have guarantee that it will be ready if needed, program is basically guarding itself from possible memory run out.

 

 

I also have 32GB of RAM but guess what my commit charge is 34GB with DCS open, so without swap file I would be no go.

 

 

 

Commit charge is a meaningless number by itself.  how much RAM is actually in-use?  That's the important figure.  if I have 32GB of RAM, and only 16 is in use, and DCS is throwing fits about RAM while it still has 16 billion bytes available, that's a DCS problem.

[ShuRugal]

@XPACT @Flappie

Just for shits and giggles, I ran the game with Procmon capturing activity, just to see if anything interesting happened. Lemme know if you want the procmon dump, it's about 4 GB, and aside from just over a million NAME NOT FOUND events, is pretty dry.  Top level counts for event results:

 

 

For shits and giggles, I also threw a paging file on my 1TB platter drive at 32GB (which is pretty fucking pointless, because that drive would take 10 minutes to swap 32 GB into RAM... faster to read it off the SSDs it actually lives on) just to see how memory allocation behaved.  Even with all this swap space to play with, DCS never requested a commit greater than 17 GB, and that my total commit never exceeded 28GB.

 

Wasn't able to reproduce crash conditions (with the Hornet, it's pretty sporadic, only about 1 crash to desktop every 10-15 hours of gameplay on average)

 

On 5/14/2021 at 6:06 PM, Flappie said:

I'm here to help you solve your DCS problem. I ask you to try something. You don't want to try. What can I do? I'm out of here.

I'm not asking for help bandaging the bad behavior in the application, I am reporting the bad behavior and providing as much data on it as I can so that ED can identify the internal engine cause of it and correct it (or not, memory allocation problems have never seemed high on their radar in the past).

Edited May 16, 2021 by ShuRugal

[jfnelson0127]

I have been away on business for the last week. Figured I'd load up a mission in the Tomcat. The first mission played through fine, but the second load of the same mission crashed to desktop as it has been doing for the past month.

 

Once again, I'm attaching my crash log. I noticed Steam completed a quick update for DCS when I started my computer. Clearly that update isn't addressing the 2nd load bug.

 

Are there any new "fixes" that have been successful for others? Deleting fxo and the various DCS folders in Temp has done nothing to solve this for me.

 

Thanks,
JN

dcs.20210516-015100.log

[YSIAD_RIP]

I had been using a specific Overclock Profile for my GPU for 1.5 years. 

• The Same GPU OC Profile was applied through several GPU Driver updates and three Windows Version updates.
  • Specifically: Windows 10 - 1809, 1909 and 20H2.

• I was always getting the odd DCS crash but after my last update to Win10 20H2 and DCS 2.7 I was getting too many crashes.
• Last week I scaled back my GPU Over Clock settings making them them less aggressive.  DCS has been much more stable with no crashes for 3 days. 

Part of me wants to put on a tin foil hat and go all conspiracy theory on how Microsoft is possibly doing stuff with their code so that CPU and GPU Overclocking is less stable and more people upgrade their computers sooner. LoL.

However; there is no way any trusted big tech company would ever Nerf their produce with updates in an effort to get customers to upgrade.   

 

Cheers

JJF. 

 

Edited May 16, 2021 by YSIAD_RIP

[Burt]

I could play other games that are more system and gpu demanding than DCS with no issues with so ever. Then run DCS open beta and crash one session after another... so yeah it’s definitely a coding thing which I know zip about.  I did all the updates-page filing- disabled all overclocking “which I shouldn’t have too” etc. This crashing issue should be top priority for us loyal customers.... I mean like before anything else is added. It’s getting worse and more aggravating each time.

[casserole]

Unfortunately, DCS keeps on crashing after today's update to version 2.7.1.6430. As a matter of fact, frequency of crashes is increasing.

I installed the new NVidia driver 466.47 to no avail. My Windows 10 Pro version is 20H2 up to date.

I include my latest crashes here below.

Have fun...

dcs.log-20210520-172706.zip dcs.log-20210520-173013.zip dcs.log-20210520-173717.zip

Edited May 20, 2021 by casserole

[casserole]

What I also noticed was that at each DCS crash, a large dmp file of about 237 MB is created in "C:\Users\UserName\AppData\Local\CrashDumps."

Maybe they are the same than the reports in "Saved Games" we are attaching here but in a much more detailed form. Don't know what to think of it.

[Flappie]

1 hour ago, casserole said:

What I also noticed was that at each DCS crash, a large dmp file of about 237 MB is created in "C:\Users\UserName\AppData\Local\CrashDumps."

Maybe they are the same than the reports in "Saved Games" we are attaching here but in a much more detailed form. Don't know what to think of it.

 

If you want to get rid of these crash dumps, try solutions listed here:

 

 

They might not work since they didn't work for Taz, but it's worth a try.

Edited May 20, 2021 by Flappie

[Fropa]

I keep getting game freezes (not responding) whenever I hop into the Tomcat in multiplayer AFTER first hopping into a different aircraft. This does NOT happen if the Tomcat is the first aircraft I hop into on multiplayer after first initializing DCS. Not sure why this keeps happening, but this problem has been consistent throughout my flying in 2.7. Log is attached. 

 

System Specs:

ASUS PRIME Z490-V ATX, ARGB

Intel i9-10900K (3.70-5.2GHz)

32GB DDR4/3200MHz Dual Channel Memory

GeForce RTX 2080 SUPER 8GB GDDR6

1TB Patriot Vi[er Gaming VP4100 Series (PCIe Gen4) NVMe M.2 SSD

Windows 10 Home 64 bit Version 20H2

HP Reverb G2 HMD

dcs.log

Edited May 21, 2021 by Fropa

[Flappie]

I've been having this Tomcat bug as well since 2.7 came out. Not 100% reproducible for me, though. I suggest you create a post in this thread with your log attached.

[C3PO]

DCS.exe.7176.rar

[Flappie]

3 minutes ago, C3PO said:

DCS.exe.7176.rar 4.39 MB · 1 download

This is a dump file. I don't read in dump files, only in chicken entrails.

[C3PO]

Ah sorry ... 

 

Had another crash just now entering co-ordinates into the UFC in the AV-8B. Entered north-ing details, pressed ENTER, and DCS crashed on me. Screen froze in VR. Flying over Syria in MP. This was on the Triple Nickle server ... I wonder if it had been updated to the latest beta as of yesterday?

dcs.log dcs.20210521-071424.crash

Edited May 21, 2021 by C3PO