Closed lozanoeg closed 2 years ago
lilkingjr1
commented
2 years ago Thank you for making this issue report. I will look into this and report back with anything I find. I will likely be working on a Yolks PR to resolve this issue, as well as clean up the egg's custom image.
lozanoeg
commented
2 years ago Thank you for making this issue report. I will look into this and report back with anything I find. I will likely be working on a Yolks PR to resolve this issue, as well as clean up the egg's custom image.
Thanks so much I'll do some of my own testing as well but I don't know the code that well so I may not be much use
lilkingjr1
commented
2 years ago I'll do some of my own testing as well but I don't know the code that well so I may not be much use
@lozanoeg The issue is mostly with the image and not the egg, so there's not much you can do on your end to try to avoid it (except for recompiling the image yourself). What I am interested in though is maybe doing some extra testing regarding your mission file loading differently in 32 and 64-bit when it shouldn't; very interesting.
lozanoeg
commented
2 years ago yeah i'll drop the mission pbo here. funnily commenting out the LDPRELOAD variable in steamcmd.sh did fix my problem. definitely a bandaid though and not a long term solution
lozanoeg
commented
2 years ago This is the mission in question co10_Escape_APEX_NATO_vs_CSAT.Tanoa.zip I had brought it up with them in the issue https://github.com/NeoArmageddon/co10_Escape/issues/116 which is where I got the thing about 64 bit being unstable
lilkingjr1
commented
2 years ago commenting out the LDPRELOAD variable in steamcmd.sh did fix my problem.
@lozanoeg What?... /home/container/steamcmd/steamcmd.sh? Which line(s) did you comment out?
lozanoeg
commented
2 years ago commenting out the LDPRELOAD variable in steamcmd.sh did fix my problem.
@lozanoeg What?...
/home/container/steamcmd/steamcmd.sh? Which line(s) did you comment out?
I commented out export LD_Library_Path and the block checking for LD_PRELOAD. On a side not is there a reason the Headless client doesn't auto connect to my server when the mission is loaded
lilkingjr1
commented
2 years ago @lozanoeg UPDATE: I haven't been able to figure out yet why that mission doesn't load properly on 64-bit, but that issue pertains more to the mission itself and it's something I might have to look at another time.
However, I have been testing running the 32-bit binary of Arma, and I think I have a solution prepared. I have been able to greatly clean up the image and remove the libnss-wrapper package, which had some code in the entry point causing issues when using the 32-bit binary of Arma. I also fixed a Pterodactyl-specific container user issue when running Arma that would crash Arma if the user wasn't properly defined (Arma requires a valid user to set up it's profiles successfully).
If you would like, you can help me test my work-in-progress image update by switching your server to use this image address:
ghcr.io/lilkingjr1/arma3-yolk:main
Regardless, I plan to open a PR in the Yolks repo soon with these changes once I perform some more tests.
lilkingjr1
commented
2 years ago On a side not is there a reason the Headless client doesn't auto connect to my server when the mission is loaded
Set "[ADVANCED] CLEAR HC PROFILES CACHE ON STARTUP" in Startup Variables to 1 and try my new image above. If it still doesn't connect, it's likely either the mission or some other config issue.
lozanoeg
commented
2 years ago If you would like, you can help me test my work-in-progress image update by switching your server to use this image address: ghcr.io/lilkingjr1/arma3-yolk:main
As of just 2 runs of the mission it seems significantly more stable i'll continue further testing and see if i can stress it out with heavier loads and other things and give any bugs I find if i find any
lilkingjr1
commented
2 years ago it seems significantly more stable
I switched it to Debian from Ubuntu and ripped out a bunch of useless packages, so maybe?...
lozanoeg
commented
2 years ago it seems significantly more stable
I switched it to Debian from Ubuntu and ripped out a bunch of useless packages, so maybe?...
Yeah i'm having more stable FPS as well and a few other benefits. headless clients work as expected and stuff like that. I am having a random crash every once in a while but i keep missing the log i'm trying to find it though
lilkingjr1
commented
2 years ago Arma crashing after the mission has successfully loaded is almost never the egg or Ptero, but feel free to post the logs if you have them (fair warning: I'm not great at reading Arma/SQF logs).
lozanoeg
commented
2 years ago it said something about profile expected = but got 'e' but either logs rolled over or I can't find the log but i'll try and see if i can get it again. i dont think it had to do with the mission
lozanoeg
commented
2 years ago @lilkingjr1 Finally repeated the output here it is
Class T1 destroyed with lock count 5
Class T2 destroyed with lock count 5
Class T3 destroyed with lock count 5
Class T4 destroyed with lock count 5
Class T5 destroyed with lock count 5
Class T6 destroyed with lock count 5
Class T7 destroyed with lock count 5
Class T8 destroyed with lock count 5
Class T9 destroyed with lock count 5
Class T10 destroyed with lock count 5
Class T11 destroyed with lock count 5
Class T12 destroyed with lock count 5
Class T13 destroyed with lock count 5
Class T14 destroyed with lock count 5
Class T15 destroyed with lock count 5
Class T16 destroyed with lock count 5
Class T17 destroyed with lock count 5
Class T18 destroyed with lock count 5
Class T19 destroyed with lock count 5
Class T20 destroyed with lock count 5
Class T21 destroyed with lock count 5
Class T22 destroyed with lock count 5
Class WeaponCloudsGun destroyed with lock count 10
Class Table destroyed with lock count 10
Class T0 destroyed with lock count 10
Class WeaponCloudsMGun destroyed with lock count 5
Class Table destroyed with lock count 5
Class T0 destroyed with lock count 5
Class All destroyed with lock count 5
Class EventHandlers destroyed with lock count 5
Class CBA_Extended_EventHandlers destroyed with lock count 5
Class MarkerLights destroyed with lock count 5
Class NVGMarkers destroyed with lock count 5
Class NVGMarker destroyed with lock count 5
Class Turrets destroyed with lock count 5
Class HeadLimits destroyed with lock count 5
Class ViewPilot destroyed with lock count 5
Class ViewCargo destroyed with lock count 5
Class ViewOptics destroyed with lock count 5
Class PilotSpec destroyed with lock count 5
Class CargoSpec destroyed with lock count 5
Class Cargo1 destroyed with lock count 5
Class TransportWeapons destroyed with lock count 5
Class TransportMagazines destroyed with lock count 5
Class SoundEnvironExt destroyed with lock count 5
Class SoundEquipment destroyed with lock count 5
Class SoundGear destroyed with lock count 5
Class SoundBreath destroyed with lock count 5
Class SoundBreathSwimming destroyed with lock count 5
Class SoundBreathInjured destroyed with lock count 5
Class SoundHitScream destroyed with lock count 5
Class SoundInjured destroyed with lock count 5
Class SoundBreathAutomatic destroyed with lock count 5
Class SoundDrown destroyed with lock count 5
Class SoundChoke destroyed with lock count 5
Class SoundRecovered destroyed with lock count 5
Class SoundBurning destroyed with lock count 5
Class PulsationSound destroyed with lock count 5
Class SoundDrowning destroyed with lock count 5
Class Reflectors destroyed with lock count 5
Class FxExplo destroyed with lock count 5
Class AnimationSources destroyed with lock count 5
Class GunFire destroyed with lock count 5
Class GunClouds destroyed with lock count 5
Class MGunClouds destroyed with lock count 5
Class Damage destroyed with lock count 5
Class DestructionEffects destroyed with lock count 5
Class Light1 destroyed with lock count 5
Class Smoke1 destroyed with lock count 5
Class Fire1 destroyed with lock count 5
Class Sparks1 destroyed with lock count 5
Class Sound destroyed with lock count 5
Class Light2 destroyed with lock count 5
Class Fire2 destroyed with lock count 5
Class Smoke1_2 destroyed with lock count 5
Class camShakeGForce destroyed with lock count 5
Class camShakeDamage destroyed with lock count 5
Class SpeechVariants destroyed with lock count 5
Class Default destroyed with lock count 5
Class LaserTarget destroyed with lock count 2
Class EventHandlers destroyed with lock count 2
Class CBA_Extended_EventHandlers destroyed with lock count 2
Class SuppressTarget destroyed with lock count 1
Class Static destroyed with lock count 2
Class DestructionEffects destroyed with lock count 2
Class EventHandlers destroyed with lock count 2
Class CBA_Extended_EventHandlers destroyed with lock count 2
Class CBA_NamespaceDummy destroyed with lock count 1
Class CBA_Extended_EventHandlers_base destroyed with lock count 5
File serverprofile/home/Player\Player.Arma3Profile, line 30: '.vol': 'l' encountered instead of '='
2:04:13 Critical:Destroying running thread!
2:04:13 Critical:Destroying running thread!
2:04:13 Critical:Destroying running thread!
2:04:13 Critical:Destroying running thread!
2:04:13 Critical:Destroying running thread!
/entrypoint.sh: line 308: 100 Segmentation fault (core dumped) ${modifiedStartup}```@lozanoeg Good catch! Could you please post the contents of your /home/container/serverprofile/home/Player/Player.Arma3Profile file (or simply the file itself)? It seems like it's either misconfigured or corrupted during runtime (rare), or maybe corrupted on shutdown (like how HC profiles tend to get corrupted).
lozanoeg
commented
2 years ago Sorry missed this here are the contents
version=1;
blood=1;
singleVoice=0;
gamma=1;
brightness=1;
activeKeys[]=
{
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Malden_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Stratis_done",
"BIS_co10_Escape_APEX_NATO_vs_CSAT.Tanoa_done",
"BIS_co10_Escape_RHS_US_vs_RU.Tanoa_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Kastellorizo_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.SefrouRamal_done",
"BIS_co10_Escape_WS_NATO_Desert_vs_SFIA_and_Tura.SefrouRamal_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Enoch_done",
"BIS_co10_Escape_RHS_US_vs_RU.Enoch_done",
"BIS_co10_Escape_RHS_US_vs_RU.rhspkl_done",
"BIS_co10_Escape_IFA3_SOV_vs_GER.tem_ihantala_done",
"BIS_co10_Escape_IFA3_UK_vs_GER.tem_anizay_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.VT5_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.VT7_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.DYA_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Fallujah_done",
"BIS_co10_Escape_SOGPF_MACV_vs_PAVN_and_VC.rhspkl_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Altis_done",
"BIS_co10_Escape_APEX_CSAT_vs_NATO.Tanoa_done"
};
soundEnableEAX=1;
soundEnableHW=0;
volumeCD=5;
volumeFX=5;
volumeSpeech=5;
volumeVoN=5;
volumeMapDucking=1;
Unfortunately this is not the same config that I got that error on but that one seems to happen randomly and not very often
lilkingjr1
commented
2 years ago Yeh, this Arma3Profile looks fine to me, but without it being in the state it was in when the server crashed, it doesn't tell me too much. However, most of the stuff in this Arma3Profile would only pertain to a client and not a dedicated server. Some of it is auto-generated, but I feel like the volume settings (what your logs say it crashed on last time) are not typically in this file for a server. So, I would edit this file to look exactly like this:
activeKeys[]=
{
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Malden_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Stratis_done",
"BIS_co10_Escape_APEX_NATO_vs_CSAT.Tanoa_done",
"BIS_co10_Escape_RHS_US_vs_RU.Tanoa_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Kastellorizo_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.SefrouRamal_done",
"BIS_co10_Escape_WS_NATO_Desert_vs_SFIA_and_Tura.SefrouRamal_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Enoch_done",
"BIS_co10_Escape_RHS_US_vs_RU.Enoch_done",
"BIS_co10_Escape_RHS_US_vs_RU.rhspkl_done",
"BIS_co10_Escape_IFA3_SOV_vs_GER.tem_ihantala_done",
"BIS_co10_Escape_IFA3_UK_vs_GER.tem_anizay_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.VT5_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.VT7_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.DYA_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Fallujah_done",
"BIS_co10_Escape_SOGPF_MACV_vs_PAVN_and_VC.rhspkl_done",
"BIS_co10_Escape_BIS_NATO_vs_CSAT.Altis_done",
"BIS_co10_Escape_APEX_CSAT_vs_NATO.Tanoa_done"
};If you don't need the save states of your Escape missions (saying that they have been completed), you could honestly make this file blank if you wanted (which would reset your Escape mission completion progress). Another thing to note: hopefully you aren't using "custom" difficulty for your missions, because it does not look to be defined in your profile.
But overall, this particular crash looks to be more like a general Arma issue than an egg issue, so hopefully when I get this yolk update done and official this issue can be closed (unless you find more to report).
lilkingjr1
commented
2 years ago @parkervcp @lozanoeg Egg update #1671 and image update parkervcp/yolks#57 should have fixed this 32-bit related issue, and this issue can likely be closed.
Panel Version
1.6.3
Wings Version
1.4.3
Service
arma3
Modified
No, I did not modify the egg
Expected Behavior
Running 32 bit binary with head clients should load the head clients
Actual Behavior
Server crashes as referenced #1412 Running 64 bit does not load missions correctly
Steps To Reproduce
Step 1 load arma3 egg Step 2 load 32 bit binary Step 3 set HC to higher than 0 Step 4 start server
64 bit logs
32 bit logs
Basically this shows up as in game a backpack on the ground not being filled in 64 bit and guards around prison dont spawn