I didn't know that -trace existed. Game froze again today, here's the last page:
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
exiting /mob/Player/Stat()
calling /mob/Player/Stat()
exiting /mob/Player/Stat()
calling /mob/Player/Stat()
exiting /mob/Player/Stat()
calling /mob/Player/Stat()
exiting /mob/Player/Stat()
calling /mob/Player/Stat()
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
I haven't seen any more refcount BUG messages in the log since the server was updated to use 1193.

However, the server did freeze twice today. Once after at least 10 hours, and again less than an hour after restarting it.
Since I am hosting about 8 servers, I wanted to know before I do anything if I can update BYOND while they are still running. My guess is I shouldn't do this. But there's always that slim chance I could be wrong and I can save myself the trouble of having to put the servers back up for these people. -- Already shutting them down and updating.

I'm having ref counts being spammed on multiple games and what brought me to this page was that one of the people I am hosting for has a game that after a few minutes it hits the maximum CPU and in the end it freezes up the game but nobody disconnects until I kill the process.

If needed, I can post traces and such.
This is still an issue for us too.
Updating while running won't work, but you should (in theory) be able to replace the exes and then call world.Reboot() in game to reconnect players.

You should always use the latest. Clearly there are some issues and when we get a block of time soon we'll work through them with you guys.
In response to Tom
I was on 1158 still. But I was thinking about replacing the .exe for the files and then rebooting. Just didn't know if it'd cause errors because on rare occasions when I do that to update byond games and I replace .dmb and .rsc files in some cases it ends up corrupting the files. Very rare, but it does. And it's not that the file didn't upload fully either. (Getting off topic here though.)
In response to Tom
After updating, the server no longer crashes. I don't see any refcounts for any games. So far so good for hosting since 4:00 AM on this new version. 11 hours straight.
All this talk about exe's is way off topic... This issue is about the Linux version of BYOND mainly DD (DreamDaemon) nothing to do with Windows........
What file type is DreamDaemon, DreamDownload, and DreamMaker then? In byond/bin/* . You know I'm on Linux as well.
Right, obviously I was talking about the Linux exes. They don't have an extension. Just always use the latest and keep reporting issues here and we'll get to the bottom of it.
Well, I don't think the crashes are an issue with BYOND entirely. I'm helping two people out with some coding. I tried tracing it and apparently this was the last few lines in the trace.txt

../libdung/server/stick.cpp:132
calling /obj/Damage/DamIcon/New()
exiting /obj/Damage/DamIcon/New()
calling /obj/Damage/DamIcon/New()
exiting /obj/Damage/DamIcon/New()
calling /mob/proc/AttackLoop()
calling /mob/proc/AttackLoop()
calling /mob/proc/GotStatusEffect()
exiting /mob/proc/GotStatusEffect()
calling /mob/proc/GotStatusEffect()
exiting /mob/proc/GotStatusEffect()
exiting /mob/proc/AttackLoop()
exiting /mob/proc/AttackLoop()
calling /obj/Anim/Anims/New()
exiting /obj/Anim/Anims/New()
calling /mob/proc/AI_AttackLoop()
exiting /mob/proc/AI_AttackLoop()
calling /mob/proc/AI_AttackLoop()
exiting /mob/proc/AI_AttackLoop()
calling /mob/proc/AI_AttackLoop()
exiting /mob/proc/AI_AttackLoop()
calling /mob/proc/AI_AttackLoop()
exiting /mob/proc/AI_AttackLoop()
calling /mob/proc/AI_AttackLoop()
exiting /mob/proc/AI_AttackLoop()
calling /mob/Stat()
../libdung/server/slinkdef.cpp:1603
../libdung/server/slinkdef.cpp:1677
exiting /mob/Stat()
../libdung/server/stick.cpp:174
../libdung/utils/timelib.cpp:244
server msg type=2,length=53
calling /turf/MouseEntered()
exiting /turf/MouseEntered()
../libdung/server/slink.cpp:351
server msg type=115,length=1908


After interviewing the owner of this game, they told me that the last thing they were doing was moving. There was only 2 players on the game at the time and they were both reporting the same thing, movement.

calling /turf/MouseEntered()
exiting /turf/MouseEntered()


I don't know if this is related to a bug in BYOND or something faulty in their codes.. but this game has so much loops it's not even worth fixing in my opinion...




The other game that crashed, the user stated that he turned on something and it caused a crash. I'll be asking him some questions and report it here. I don't know if this information helps, but who knows. Someone else may run in to this problem someday. I know my games don't. <-- (I questioned this one. It was just poor coding. He turned on some electricity thing in-game and it caused something to spiral out of control. I guess there wasn't a limit on how fast something was being called.)

Edit: This trace was done after I updated BYOND. The game was up for 11 hours without any players online and as soon as I tell them to log on and try playing the game to see what happens, 5 minutes pass and it crashes.
I wish the game was crashing. Then I'd have something to look at in the log or something. Unfortunately, it's just freezing and forcing me to kill it manually, which doesn't affect the log.
Just as an additional note, the refcount fix that went in recently affects obounds() and would impact objs and mobs. Those are the 3:nnn and 2:nnn entries you see with the refcount bugs.

The obounds() fix does not impact clients references (5:nnn). However, when I added that fix I did find that several skin routines were malfunctioning if done before the skin was ready, like in client/New(), and that could have caused some issues including the proc stack getting messed up and the refcounts being off.
In response to Kaiochao
My games don't crash either. They just hit the peak of the CPU and sit there. Then I have to kill it my hand.
Froze again.

exiting /atom/Read()
../libdung/server/sfile.cpp:1523
../libdung/server/sfile.cpp:1529
../libdung/server/sfile.cpp:1534
../libdung/server/sfile.cpp:1537
../libdung/server/sfile.cpp:2792
../libdung/server/sfile.cpp:2807
../libdung/server/sfile.cpp:1523
../libdung/server/sfile.cpp:1529
../libdung/server/sfile.cpp:1534
../libdung/server/sfile.cpp:1537
../libdung/server/sfile.cpp:2810
../libdung/server/sfile.cpp:1213
../libdung/server/sfile.cpp:1215
../libdung/server/sfile.cpp:1217
../libdung/server/sfile.cpp:1222
../libdung/server/sfile.cpp:2813
../libdung/server/sfile.cpp:2816
../libdung/server/sfile.cpp:2818
../libdung/server/sfile.cpp:2822
../libdung/server/sfile.cpp:2825
../libdung/server/sfile.cpp:2827
calling /atom/movable/PM/New()
exiting /atom/movable/PM/New()
calling /atom/Read()
../libdung/server/sfile.cpp:2792
../libdung/server/sfile.cpp:2807
../libdung/server/sfile.cpp:1523
../libdung/server/sfile.cpp:1529
../libdung/server/sfile.cpp:1534
../libdung/server/sfile.cpp:1537
../libdung/server/sfile.cpp:2810
../libdung/server/sfile.cpp:1213
../libdung/server/sfile.cpp:1215
../libdung/server/sfile.cpp:1217
../libdung/server/sfile.cpp:1222
../libdung/server/sfile.cpp:2813
../libdung/server/sfile.cpp:2816
I've been hosting a Hazordhu server for nearly 24 hours without any kind of freeze or crash on Windows 8 Pro.
I should probably also mention that it only started freezing after updating the BYOND version the game was compiled in. It used to run fine on the previous DD version, 498.1158, until the game itself was updated.
In response to Kaiochao
Kaiochao wrote:
I should probably also mention that it only started freezing after updating the BYOND version the game was compiled in. It used to run fine on the previous DD version, 498.1158, until the game itself was updated.

My game is compiled in 498, though the issue didn't seem to start until around the time that 499 was released. Possibly something to do with the clients? Seems unlikely.
In response to Murrawhip
I bet it does have something to do with older clients, but I have nothing to go on.
In response to Tom
Should I continue posting traces? They're not the same all the time.
Page: 1 2 3 4 ... 6 7 8