Jump to content

Consistent Crashes with Update

cnoblenyc

By cnoblenyc
in Game Crash

 Share

Recommended Posts

cnoblenyc

cnoblenyc

Posted
Posted (edited)

Morning All,

I've had a very stable experience for well over a year up to the latest update.  As soon as the update to the latest version occurred, I cannot complete a single flight in F/A-18 or F-15E (F-15E won't even load to start the flight...) To be clear, couple weeks ago (before the update) everything was stable.  I have re-installed DCS completely, deleted "saved games" data, reinstalled the nvidia drivers... to no avail.  I've tried both "DCS" and "DCS-MT".  I usually just CTD, but every once in a while, I get a "proper dump".

Attached is the dump logs from the last "proper dump" and the latest log of a CTD.

 

 UPDATE:  Added 2 more logs - 1 "proper crash" and one direct CTD.  Bothe with F/A-18.. didn't even start the flight.dcs.log-20240524-133144.zip 

I love this simulator, thanks for all the hard work behind it.  Cannot wait to see where it goes next!

dcs CTD.log dcs.log-20240524-023442.zip dcs w-dump.log

dcs - F-18 - CTD Before flight started.log dcs - F-18 - Proper crash - on mission load.log dcs.log-20240524-133144.zip

Edited by cnoblenyc
Updated
cnoblenyc

cnoblenyc

Posted
  • Author
Posted

I spoke too soon... had a couple longer flights (was actually able to do a trap once). I've tried everything that I can think of and all the options I can find on the forum. Including going back to 2.8.  I've replaced the driver and used DDU to do a complete clean install of Nvidia. After several more hours of tinkering and re-installing and adjusting, I cannot get more than half a flight in.. at best. It usually fails when loading the missions with "World Pre-load". 

I've attached a few more dumps and logs...

It's absolutely unplayable at this point - hope someone can help.

1st run with dcs 2.8.7.4271 dcs.log 1st run with dcs MT2.8.7.4271.log dcs.log-20240525-185541.zip 2nd run with dcs MT 2.8.7.4271.log dcs.log dcs.log.old

hornblower793

hornblower793

Posted
Posted

What sort of mission are you trying to run? - the log I looked at generated a crash trying to generate the mission briefing

Windows 11 Home ¦ Z790 AORUS Elite AX motherboard ¦ i7-13700K ¦ 64GB Corsair Vengeance DDR5 memory @ 5600MHz ¦ Samsung 990 Pro 1TB SSD for OS, Samsung 980 Pro 2TB SSD for DCS ¦ MSI GeForce RTX 4090 Gaming X Trio 24GB ¦ Virpil WarBRD base with VFX grip, Thrustmaster A10c and F/A-18 grips ¦ VKB Gunfighter Mk4 and MCG Pro ¦ Thrustmaster Warthog Throttle ¦ VKB STECS Throttle ¦ Virpil TCS rotor base with Shark and AH-64D  grips ¦ MFG Crosswinds ¦ Total Controls Multi-Function Button Box ¦ Pimax Crystal

cnoblenyc

cnoblenyc

Posted
  • Author
Posted

Hi Hornblower - 

ANY mission - pre-made or self generated.  I've generated my own simple mission (Carrier and F/A-18) attached hereto.  I CTD 80% of the time when it gets to when it hits "World Preload 40+/60".   I've attached the corresponding log.  The log has multiple errors with inability to open files...

CATI - Training.miz dcs.log.old

Flappie

Flappie

Posted
Posted

You may have an issue with the drive DCS is installed on. 

If you have enough room on another drive, just copy the entire DCS World folder to it, and launch it (/bin-mt/DCS.exe). Then see if it crashes.

 

---

Flappie

Flappie

Posted
Posted (edited)

I see your pagefile is really tiny: 2 GB. 

2024-05-27 23:46:28.655 INFO    APP (Main): CPU cores: 8, threads: 16, System RAM: 32684 MB, Pagefile: 2048 MB

If you have a fast drive with a lot of free space, you should either let Windows handle the pagefile size, or manually set a bigger pagefile yourself (init size: 32768, max size: 32768 - which means a 32 GB pagefile).

You should also reduce the sim's memory needs by lowering the Preload Radius option. It's currently set to the max (150000), which is a bad idea as long as you don't have 64 or 128 GB of RAM.

Edited by Flappie

---

Flappie

Flappie

Posted
Posted

This crash log is interesting: 

2024-05-29 04:15:34.131 INFO    EDCORE (11376): try to write dump information
2024-05-29 04:15:34.135 INFO    EDCORE (11376): # -------------- 20240529-041534 --------------
2024-05-29 04:15:34.135 INFO    EDCORE (11376): DCS/2.9.5.55300 (x86_64; MT; Windows NT 10.0.22631)
2024-05-29 04:15:34.135 INFO    EDCORE (11376): C:\WINDOWS\SYSTEM32\ntdll.dll
2024-05-29 04:15:34.135 INFO    EDCORE (11376): # C0000005 ACCESS_VIOLATION at 00007ffc2a391ed4 5A:000346B0
2024-05-29 04:15:34.135 INFO    EDCORE (11376): SymInit: Symbol-SearchPath: 'C:\Program Files\Eagle Dynamics\DCS World\bin-mt;', symOptions: 532, UserName: 'cnobl'
2024-05-29 04:15:34.136 INFO    EDCORE (11376): OS-Version: 10.0.22631 () 0x100-0x1
2024-05-29 04:15:34.136 INFO    EDCORE (11376): 0x0000000000041ed4 (ntdll): RtlReAllocateHeap + 0x284
2024-05-29 04:15:34.136 INFO    EDCORE (11376): 0x0000000000041caa (ntdll): RtlReAllocateHeap + 0x5A
2024-05-29 04:15:34.136 INFO    EDCORE (11376): 0x00000000000118e9 (ucrtbase): _realloc_base + 0x39
2024-05-29 04:15:34.136 INFO    EDCORE (11376): 0x0000000000e7445b (DCS): SW + 0x9B33BB
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000d0355e (DCS): SW + 0x8424BE
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000d04108 (DCS): SW + 0x843068
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000d03b60 (DCS): SW + 0x842AC0
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000d4d56e (DCS): SW + 0x88C4CE
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000d2948d (DCS): SW + 0x8683ED
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000d3d116 (DCS): SW + 0x87C076
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000d6cfeb (DCS): SW + 0x8ABF4B
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000de8b10 (DCS): SW + 0x927A70
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000de7b5f (DCS): SW + 0x926ABF
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000d0b9c1 (DCS): SW + 0x84A921
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000d38a89 (DCS): SW + 0x8779E9
2024-05-29 04:15:34.137 INFO    EDCORE (11376): 0x0000000000dc62be (DCS): SW + 0x90521E
2024-05-29 04:15:34.138 INFO    EDCORE (11376): 0x0000000000d0b9c1 (DCS): SW + 0x84A921
2024-05-29 04:15:34.138 INFO    EDCORE (11376): 0x0000000000d19264 (DCS): SW + 0x8581C4
2024-05-29 04:15:34.138 INFO    EDCORE (11376): 0x0000000000d19562 (DCS): SW + 0x8584C2
2024-05-29 04:15:34.138 INFO    EDCORE (11376): 0x0000000000e75346 (DCS): SW + 0x9B42A6
2024-05-29 04:15:34.138 INFO    EDCORE (11376): 0x000000000001257d (KERNEL32): BaseThreadInitThunk + 0x1D
2024-05-29 04:15:34.324 INFO    EDCORE (11376): Minidump created.

 

The "RtlReAllocateHeap" function is from Windows. It is used to handle things into memory.

You can check your RAM using MemTest86:

 

---

Flappie

Flappie

Posted
Posted

Please try reinstalling Visual C++ redistributables.

  1. Browse to "./DCS World/distr" (or "./DCS World OpenBeta/distr").
  2. Run redist_x64.exe and click "Repair".
  3. Now run vc_redist.x64.exe and click "Install". If the installation comes to a halt, it's OK: it means this package is already installed on your computer.

---

cnoblenyc

cnoblenyc

Posted
  • Author
Posted

No BSD's at all.  Always goes CTD. In fact, I rarely get just a CTD. 9 out of 10 times, it crashes properly (crash-dump file created).

BIOS is up-to-date (latest version from Asus) as are the Intel chipsets.

 

cnoblenyc

cnoblenyc

Posted
  • Author
Posted

APC Battery backup... bypassed - No change in behavior.

I replaced the vid card - new Asus 4070 ROG.  Same issues.

I reverted back to  2.8.7.42718 - Same issues.  

I'm fairly sure now the issue is on my end - the only thing left to try seems to be replacing the MB which is not fun.

The challenge I have is this is the ONLY piece of software that gives me issues.  MSFS operates or hours at a time without any issue... and it pushes the rig hard.

 

Any other ideas before going down that rabbit whole?

 

 

cnoblenyc

cnoblenyc

Posted
  • Author
Posted

I'd suspect if it were a faulty PCU, the system and/or other apps would have issues.  It's the Corsair AX850i  MSFS taxes the system significantly and that is stable for hours at a time...

The consistency in the logs is "# C0000005 ACCESS_VIOLATION".  What cause only DCS to to have this issue that no other application seems to have...??  What is it violating?

 

I've attached 2 screenshots from iCUE.  As you can see, the power consumption (wattage and current) goes up as should be expected, but the voltages stay rock solid across the unit.  Sadly, the only part of the system that isn't monitorable (or hasn't been changed) in the MB and CPU.. and again, my challenge is NO other application (including power-consuming ones like MSFS) cause any issues.

Spent a lot of money on this game - built the whole rig around this - to not be able to use it... getting very frustrating ,to say the least...

 

But I do appreciate your help.  Attached is 2 of the latest logs as well.

 

At crash.png

Before DCS Launch .png

dcs.log-20240601-175637.zip dcs.log-20240601-170423.zip

Flappie

Flappie

Posted
Posted
1 hour ago, cnoblenyc said:

I'd suspect if it were a faulty PCU, the system and/or other apps would have issues.

👇

On 12/3/2020 at 7:18 PM, thinkerdreamer said:

Since I don't crash in any other game and since my computer never randomly crashes when not gaming, I think it's unlikely to be the PSU. I don't have SSLR enabled. My RAM is not overclocked.

A few moments later... (actually 2 months)

On 2/13/2021 at 4:44 PM, thinkerdreamer said:

@Flappie Hey, late update but I just finished the final check and it seems to be working perfectly now. My PC had a 400W peak and 300W continuous PSU. I swapped it with a 750W PSU and that seems to have fixed it. I never thought that could be the problem, but I didn't realize it was such a <profanity>ty PSU in there. Thanks for your help.

1 hour ago, cnoblenyc said:

The consistency in the logs is "# C0000005 ACCESS_VIOLATION".  What cause only DCS to to have this issue that no other application seems to have...??  What is it violating?

This crash code is the most common in DCS. Actually, it is so common that it does not mean much.

You should check Windows Event Viewer (System and Apps logs) at the time of your DCS crashes. You may find interesting messages..

---

FusRoPotato

FusRoPotato

Posted
Posted

Seems like I and two friends are also seeing this, though I have only seen it once and they see it as often as once per hour. Commonality is C0000005 and resource exhaustion warning, ran out of virtual memory, consumed 30 gigs of ram and 51 gigs of virtual memory.

This is another memory leak.

FusRoPotato

FusRoPotato

Posted
Posted (edited)
12 hours ago, Flappie said:

A memory leak is reproducible.

A memory leak into virtual memory was reproduced, albeit, a slow one that took a long time for me to monitor.

While I haven't datalogged the PC's of the other individuals, they have the same events on record and I am seeing a very slow and gradual climb of virtual memory allocation for this process over time through perfmon, along with some unusual behavior I don't find in other games. After approximately 2 hours on a multiplayer Syria server, resting in an FC3 airfraft, I often saw ram usage floating on average between 12-17 gigs. I also selectively limited my graphics to low textures and relatively low details allowing my 7900XTX to average usage around 11 gigs Vram for 1440p rendering. However, Private Bytes for this app slowly climbed over time to a point of reaching 41 gigs whilst starting at 22 gigs and was incremented by cycling F2 on aircraft sprinkled across the larger map of Syria. The odd thing I noticed was a high amount of shared GPU Process Memory of 2.5 gigs while using 11 actively on a 24 gig card.

When using high detail textures and higher settings (but still 1440p), the game initializes loading into a multiplayer server with 32 gigs virtual mem fully allocated before slotting in (as an increase from a point in time prior to loading the game), but 17 gigs ram used. From here, slotting in makes ground textures take more than a minute to load in (invisible ground). Vram usage floats between 18-19 gigs. When using F2 to cycling around between aircraft scattered around Syria, the Private Bytes appears to grow at a much faster rate. After about 30 minutes of cycling between F2 views, I hit the cap of my virtual memory (a low cap automatically set to 52 gigs), received the resource warning event, and earned a new C0000005 in the log file.

Some takeaways:

  1. Lower settings seem to cause the problem to grow more slowly, possibly allowing the problem to go away.
  2. Virtual memory limits can simply be increased, also extending the time until C0000005
  3. There are still memory management issues like perhaps over-committing to page using deprecated data the game no longer needs, then an updated garbage collection method skips those deprecated objects because it doesn't expect them. I'm not seeing cases where the game uses more than 17 gigs of ram, so why is virtual memory climbing away? Maybe there's a relationship between this and some of the new desync issues.

I found a third person in my smaller discord of goons who also ran into this problem and decided to just walk away from it. They refused to confirm the presence of similar events. I'm getting the impression people are generally very reluctant to report issues like this for whatever reason, as it was difficult to convince the others to dig into their events.

@cnoblenyc FYI, windows key -> Event Viewer -> Windows Logs -> System -> right-hand-side Filter Current Log -> Check Warnings and apply. Look for source types Resource-Exhaustion-Detector. If you see something like that, try increasing your virtual memory using the guide pinned at the top of this forum.

Edited by FusRoPotato
 Share
Go to topic listing

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...