DCS freezes/unresponsive at approximately 90 minutes mission elapsed time

[Q-Ball]

I know there have been some other crashes with similar behavior around the Tomcat, and normally my squadron files with mods. Last night we used a mission with no mods, so I thought I'd submit some track files for review. It seems that when the player's client crashes, there is no proper track file recorded. However, I have included the track from the server, a track from my RIO and a track from a squadron-mate who was also flying with us last night. In the track, the issue should occur some time between 08:21 and 08:22L (mission begins at 06:48L), or approximately 93-94 minutes of mission-elapsed time. My RIO described the issue as, "The aircraft just came to a stop in mid-air," and my review of the server track shows the aircraft slowing to 8kts IAS. Others in the server described aircraft as "frozen in mid-air" in the pattern around The Boat.

 

My aircraft in all tracks will be F-14A-135-GR(Covergirl VF-211 |307|)

 

Bug: DCS freeze/unresponsive after 60-90+ minutes of mission time. Only recourse is to kill in task manager.

Can I reproduce it 100%: no

 

If not 100%, how often out of 10: Roughly 70% of missions that last longer than an hour. This freeze/crash issue cropped up in the OB sometime earlier this year.

 

How to reproduce/ description: This crash has occurred in multiple situations, but most often it happens in the vicinity of the carrier (I am not sure if it's a function of mission elapsed time or not, but there is a correlation).

DCS Version: OpenBeta 2.7.16.28157

System Specs:

CPU i7-4770k (non-OC)

GPU AMD Vega 64

32GB RAM

SSD OCZ 2TB

OS: Win10

Peripherals:

Joystick: TM Warthog

Throttle: WinWing SuperTaurus

Pedals: MFG Crosswind

Others: Winwing Takeoff Panel

 

Headtracker: TrackIR

Mission File: attached

Log: attached

Server log: attached

 

Track: 

My track (Appears to just be a mission file copy): Dropbox - Track file

Server track: Server track file

RIO track: RIO Track File

Squadron-mate track: Squadron-mate track file

 

Video/ Screenshots: Did not capture

 

Mods: We disabled all of our mods, and this mission was created without mods.

dcs.log FNFN-faround.miz dcs_server-20220819.log

Edited August 20, 2022 by Q-Ball
Better description of crash behavior in title

[Q-Ball]

Crash behavior like the one reported in this thread:

https://forum.dcs.world/topic/303893-freezing-a-lot-log-attached/#comment-5006887

[dggoofy]

I had this issue too. But nothing to do with timing. I had freeze and crash after 15 or 90 minutes, but usually about 50 NM away from the carrier.

I did 2 times the same mission : first time : crash. 2nd : no issue.

The only thing I did the 2nd time was : NOT using the F2 key view! It seems strange, but I could end the campaign without any problem just by avoiding to press the F2 view key!

Test it...

[Q-Ball]

5 hours ago, dggoofy said:

I had this issue too. But nothing to do with timing. I had freeze and crash after 15 or 90 minutes, but usually about 50 NM away from the carrier.

I did 2 times the same mission : first time : crash. 2nd : no issue.

The only thing I did the 2nd time was : NOT using the F2 key view! It seems strange, but I could end the campaign without any problem just by avoiding to press the F2 view key!

Test it...

Was this SP or MP? Edited to add: I am not sure I've ever run SP long enough recently to notice, but all of my problems with this freeze/crash behavior have been in MP.

Edited August 21, 2022 by Q-Ball
Rephrasing question

[Q-Ball]

A bump on this and also linking to a thread over in the DCS 2.7 bug reporting with similar issues. I don't know if track files other than my own client file will be helpful, but we (my squadron) are happy to try to keep reproducing this and assist any way we can. However, we are not flying with F-14 AI.

 

[IronMike]

3 minutes ago, Q-Ball said:

A bump on this and also linking to a thread over in the DCS 2.7 bug reporting with similar issues. I don't know if track files other than my own client file will be helpful, but we (my squadron) are happy to try to keep reproducing this and assist any way we can. However, we are not flying with F-14 AI.

 

Does this mean you have no AI at all in your mission? And yes, a log and track would help tons if possible. However, tracks need to be short, which makes that a problem. A 90 minute track can lead to pretty much anything unfortunately.

[Q-Ball]

On 8/24/2022 at 1:29 PM, IronMike said:

Does this mean you have no AI at all in your mission? And yes, a log and track would help tons if possible. However, tracks need to be short, which makes that a problem. A 90 minute track can lead to pretty much anything unfortunately.

In the original mission, we did not use any AI. However this past Friday we created a mission (attached) specifically intended to force a crash of the clients. Tomcat AI, tankers, and two CVs. All four Tomcat pilots crashed (only my human RIO did not). This is in contrast to our previous test attempts, where at most we saw the crash of a single client. However, the first crash did not happen until about 45 minutes in. This seems like something that consistently happens later in a mission (in our experience), for whatever reason.

I know you said that large track files are a pain, but I'm including what I have, in the hope that it's useful. Of course, none of the client crashes produce a track file.

Server Track

RIO Track

In_Crash_We_Trust.miz

Edited August 29, 2022 by Q-Ball
Clarity

[Callsign JoNay]

Did all the Tomcat pilots crash at the same time? Or within 30 seconds or so of each other?

I don't think it's related to Tomcat AI. The most common place I have these freezes are on the Hoggit training server, and AFAIK there are no AI Tomcats flying around. Lots of tankers, AWACS, and L-39s flying around for target drone practice though.

[zZzMan]

On 8/29/2022 at 2:40 PM, Callsign JoNay said:

Did all the Tomcat pilots crash at the same time? Or within 30 seconds or so of each other?

I don't think it's related to Tomcat AI. The most common place I have these freezes are on the Hoggit training server, and AFAIK there are no AI Tomcats flying around. Lots of tankers, AWACS, and L-39s flying around for target drone practice though.

I am a squadron mate of Qball. You and I have been in another thread for a similar reason. I beleive there are 2 different bugs at play here. One that causes a crash consistently about 60 to 90 min in per client. Out of 12... one person would crash. The other bug is the F-14 AI issue. This bug presents more like yours. A crash later into a mission [iirc our last test I crashed at 42 min] where those who are close to each other crash together. This second condition [mission file with both active and inactive F-14 AI plus other items] has shown a 12/12 crash rate. Crash 1 seems to be an ED issue. Crash 2 seems to be a Heatblur issue. @IronMike just @ you for reference. [Sorry for formating, on mobile]

[zZzMan]

To add this thread was originally made in relation to Qball and I thinking this was the same bug. Last week our tests showed that the assumption caused us to exclude data outside of the F-14 airframe. To make this clear this thread is for a crash that presents differently then the F-14 AI bug as we have seen F-18s crash singularly aswell. [Referencing above] Crash 1 affects other aiframes. Crash 2 [AI] only affects the 14 so far in testing. I could be wrong, I have minimal experience in programming and QA testing. 

Edited August 31, 2022 by zZzMan
Reworking Grammer to make some sense

[Q-Ball]

I don't think the proposed fix solved the issue. We re-ran "In_Crash_We_Trust.miz" again last Friday and all but one Tomcat (out of six) crashed.