- User Since
- Sep 28 2013, 7:01 PM (329 w, 6 d)
May 10 2016
I observed that problem to, in some cases it seems to happens when particles are shown. Could also be memory related, some players reported that they fixed the problem with switching memory manager.
No more crashes with PERF 4, In fact stability for linux binaries have never been that high
Nope issue is still present;
I am uploading only relevent core dump crashes log (related to libc6) to that issue on PERF 2
Filename : relevent_core_dumps3.zip
Thank you for your time
Also worth mentionning that the server is purelly vanilla, no mods or extension calls.
first one since the hotfix
sure, there you go.
Most crashes are reported @
0x945a250: mov 0x60(%eax),%eax
while EAX is null:
eax 0x0 0
Another crash is reported @
> 0xf75e1b91: movdqu %xmm0,(%esi)
Do you simply realise that no (bidmp + mdmp) are generated by linux server.
The crash dumps should be more that enougth to proceed to a debugging. (the execution stack point and the memory dump are present in it
It would be nice to have the ticket read before getting a copy paste anwser and categorized into "need more info".
Putting some more basinc imformations on the core dumps
Also might be related to http://feedback.arma3.com/view.php?id=23593
I have uploaded some RPTs, but sadly and for some unknown reasons the process crashes before being able to generate crashdumps. As you can see the last lines of the rpts are corresponding to the first 2 line of the crash report. Crash report worked before the update to 1.30 but is not working anymore.
Also when you connect to an UAV and this uav get destroyed, for the server you are still in the possision you were when you connected or where you were when the uav got destroyed.
This way you current possition is not sent back to the other player, making you invisible to them, and preventing any server script based on your position to be run correctly.
Once you enter a vehicle the server seems to update the position of the player.
After several tests it seems that you can be stuck in any kind of vehicules, but it is harder to reproduce.
It seems that those crashes are happening when a player connect or disconnect since the last line in the server console log is usually one of client connection /disconnection and usually not complete since the server crash before the buffer is written in the server consile log file eg:"19:13:54 Player XXXXXX connected (id=00000000" (this line missing ")."
Categorised as random, so no reproducibility
idk, sometime that the case, there are my server logs line when the crash happened:
11:19:02 Player (REDACTED) disconnected.
11:19:011:44:52 BattlEye Server: Initialized (v1.187)
11:44:52 Host identity created.
13:55:01 BattlEye Server: Player #21 (REDACTED)((REDACTED)) connected
13:55:01 Player Book14:05:32 BattlEye Server: Initialized (v1.187)
14:05:32 Host identity created.
17:03:24 Player (REDACTED) connecting.
17:03:25 BattlEye Server: Player #21 (REDACTED) - GUID: (REDACTED) (unverified)
17:03:26 BattlEye Server: RCon admin #1: (Global) BEC : Admin (REDACTED) Connected
17:03:17:12:05 BattlEye Server: Initialized (v1.187)
18:47:12 BattlEye Server: (Side) (REDACTED) : oh :D
18:47:23 BattlEye S19:05:01 BattlEye Server: Initialized (v1.187)
19:05:01 Host identity created.
Maybe this issue should be closed since the hotfix has been pushed to stable?
If you look at http://arma3.swec.se/game/list , you wil see that not many games with heavy AI presence (domination, becti,...) cannot get passed the 2 hours mission time.
I also have this issue on both offroad which inherit from the 'I_G_Offroad_01_F' class. The JIP process is unable to get the global variables assigned to these vehicles.
I am on stable, server on PERF 3 binaries (1.18.124360)
same here, nice spotting
Same here, strange things I noticed:
- Happens when a vehicule takes damage
- Affect one side (the one which vehicule gets damage) more than the other