The 2.7. crashes - some conclusions

[wernst]

What we know so far:

 

• the crashes cannot be reproduced consistently 
• the ED devs obviously have no crashes
• there are many possible solutions but none last longer than a few days
• A lot of standard recommendations: OS update, graphics driver update, no mod, no OC, optimize page file settings, clean some folder content    and . . . and. . . more excotic ones, but none seems to work truly.
• confusing OS recommendations: Win10 should be up-to-date, but not too up-to-date, but ultimately only Win 8 comp mode is the most promising workaround for the meantime.

 

The main problem is that any attempt to track down the main fault is made difficult by the fact that the disappearance and appearance of the crash comes completely random in time. It seems that there is a trigger in the program code or in the OS environment which strikes randomly,

I call it „trouble trigger“. Perhaps this trouble trigger is not in the program code itself but maybe in Windows 10.

 

The devs may have a particular clean OS which is free from any unwanted trigger. It is their reference system, which is needed for coding in a well known and structured environment. That would explain why the devs did not notice any abnormalities when changing from 2.5.6 to 2.7.xxxx. And it could explain why the Win 8 compatible mode is the best workaround so far as it is free from any DCS injurious triggers. 

 

I don't think the devs provided  a badly coded 2.7. open beta. Every beta version before was more or less stable, fully usable with only minor cosmetic restrictions, no fatal crashes at all. The devs made the 2.7. open beta in their own, dedicated OS environment, which seems to be  a kind of „steril“, which almost nobody of us has.

Edited May 27, 2021 by wernst

[Ironhand]

3 hours ago, wernst said:

...The main problem is that any attempt to track down the main fault is made difficult by the fact that the disappearance and appearance of the crash comes completely random in time...

 

This is the most troubling thing about this bug. I wonder if there’s anyone with this bug for whom none of the temporary fixes have worked and for whom every second mission crashes. If so, that’s who the devs need to be working with.

[Baz000]

I'm getting reproducible hard crash to desktop just observing the F-16 instant action mission called "Fighter Sweep" and the DCS logs just get cut off so DCS is not reporting why it is experiencing the crash.

 

All I do is let my plane get shot down and then watch the AI play the mission and cycle the views using F6, F2, and F3 cameras until there is no more fighting left.

 

 

dcs.log

Edited May 26, 2021 by Baz000

[Ironhand]

And you’ve volunteered, I see. Good.

Edited May 26, 2021 by Ironhand

[zork824]

The only consistent solution is to switch DCS to Win 8 compatibility mode.

[Guitar_Shredder]

On 5/26/2021 at 2:27 PM, zork824 said:

The only consistent solution is to switch DCS to Win 8 compatibility mode.

I been back and forth with this and Win 8.1 is more reliable now.. Every time I switch back, it crashes.. I flew multiple missions in one sortie that was 3 hours long and it CTD when it wasnt Win 8.1...   I hope they fix it.

[Flappie]

Hi @Guitar_Shredder. This issue was fixed. Îs your game up-to-date?

Please attach your dcs.log file ( "C:\Users\username\Saved Games\DCS...\Logs\dcs.log" - may appear as "dcs" ).

Edited June 6, 2021 by Flappie

[Slice313]

I though the issue was fixed but I started getting crashes back on second mission load. I repaired the game  (slow method) and deleted the FXO folder, but still crashes (less than before but sometimes crashes)

 

Also, now, crash logs and dumps are never created.

 

Pd: I was flying the Tomcat

[Flappie]

The Tomcat seems to have a different issue. It's actually a freeze, not a crash, hence no complete logs and no dump.

See here:

 

 

[Burt]

Every once in awhile I’ll get that Tomcat freeze too. But not all the time. In a multiplayer, not that often as single player.