kama a écrit :
On Wed, 23 Nov 2005, Vincent H. wrote:
Hi,
Same choke problems with some players and I also noticed instability
about fps servers after this update.
We run a Couter-strike 1.6 decicated linux Server, some Details:
[*] Amd Sempron 2500+ ;
[*] 512 ram ;
[*] OS: Debian Sarge ;
[*] Kernel : Linux 2.6.8-2-386
[*] Glibc : Release stable, Version 2.3.5
[*] 18 slot Slots ;
Also:
[*] Metamod 1.18p26 ;
[*] HLGuards 1.18 & SteamBan activated
Once patch applied, the server crashes randomly very often with the
following error :
----------------------------------------------
CRASH: mer nov 23 01:40:49 CET 2005
Start Line: ./hlds_amd -game cstrike -debug -port 27015 +maxplayers 18
+map de_duztec2 +ip 193.24.215.36 -pidfile hlds.1681.pid
Using host libthread_db library "/lib/tls/libthread_db.so.1".
Core was generated by `./hlds_amd -game cstrike -debug -port 27015
+maxplayers 18 +map de_duztec2 +ip'.
Program terminated with signal 11, Segmentation fault.
#0 0x400c4cc3 in ?? ()
#0 0x400c4cc3 in ?? ()
No symbol table info available.
End of crash report
----------------------------------------------
So, I hope to see new update because of this crashes are really boring
:-( :-( .
Regarding these crashes. Can you see whats the last lines are in the
console?
The most common crashes occur when someone connects, when someone drops or
reconnect. But I have also got it randomly. The last strace came up with
this on the latest line at the console.
L 11/23/2005 - 14:04:41: "CriM<2><STEAM_0:A:XXXXXXXX><TERRORIST>"
triggered "Planted_The_Bomb"
You're right.
L 11/23/2005 - 13:15:28:
"<((!!))sephiroth((!!))><15><STEAM_ID_PENDING><>" connected, address
"xx.xxx.xx.xx:27005"
L 11/23/2005 - 13:15:29:
"<((!!))sephiroth((!!))><15><STEAM_0:1:xxxxxxx><>" STEAM USERID validated
L 11/23/2005 - 13:15:36: "oow<7><STEAM_0:1:xxxxxxx><TERRORIST>" killed
"~^~ Arkange ~^~<3><STEAM_0:0:xxxxxxx><CT>" with "mp5navy"
L 11/23/2005 - 13:15:42: "oow<7><STEAM_0:1:
It crashes randomly when a player connect / reconnect / drops. Also
randomly on change map.
I will try to run the server under gdb and try to reproduce the bug .And
i will post a
backtrace of the segfault error.
So... let's go. ^^
The crashes I get is actually kill() that are sent from the server. So no
coredump is made. (This strace was made by previous hlds-version)
strace:
- snip -
read(1, "\203", 0) = 1
gethostid?() = 0x1
read(1, "\203", 0) = 1
gethostid?() = 0x1
read(1, "\203", 0) = 1
SYS_175() = 0
mincore(0xbfbfeadc, 0, []) = 0
kill(93381, 32) = 0
mincore(0xbfbfe2a8, 0, []) = 0
SYS_175() = 0
SYS_175() = 0
mincore(0xbfbfe098, 0, []) = 0
getdomainname( <unfinished ...>
- snap -
This is one just created:
- snip -
mincore(0xbfbfea4c, 0, []) = 0
mincore(0xbfbfea4c, 0, []) = 0
mincore(0xbfbfea4c, 0, []) = 0
mincore(0xbfbfd9bc, 0, 0) = 0
getdomainname("ÉÃ�U‰åƒ‹E
…ÀtèÚhôÿÇ", 3217025832) = 0
gethostid?() = 0
mincore(0xbfbfeadc, 0, []) = 0
mincore(0xbfbfea4c, 0, []) = 0
recv?(674183512, 0x4, 3217021340,
MSG_PEEK|MSG_CTRUNC|MSG_EOR|MSG_DONTWAIT|0xf00) = 16
recv?(674183512, 0x4, 3217021340,
MSG_PEEK|MSG_CTRUNC|MSG_EOR|MSG_DONTWAIT|0xf00) = 23
fchdir(0) = 1132751305
kill(24956, 32) = 0
mincore(0xbfbfd328, 0, []) = 0
SYS_175() = 0
SYS_175() = 0
mincore(0xbfbfd100, 0, []) = 0
getdomainname("ƒÄ…Àu¤Qj", 3217019128) = 0
SYS_175() = 0
fstatfs(135975268, {f_type=0, f_bsize=1, f_blocks=0, f_bfree=0,
f_bavail=0, f_files=0, f_ffree=0, f_fsid={41, 136395624}}) = 0
openbsd_poll(0x1, 0x1, 0x2bd2dc97 <unfinished ...>
- snap -
notice 'kill(24956, 32)'. 32 is SIGTHR (thread interupt) and are sent to
pid 24956 which probably is a thread that hlds have started since the
parent pid was 24954.
I am also seeing a lot of these errors in the strace:
recv?(674183512, 0x4, 3217021340,
MSG_PEEK|MSG_CTRUNC|MSG_EOR|MSG_DONTWAIT|0xf00) = -1 E??? (errno -11)
-11 should be ( if I understand correctly ):
[EAGAIN] The socket is marked non-blocking, and the receive operation
would block, or a receive timeout had been set, and the timeout expired
before data were received.
But thats a legacy call. in errno.h it is specified:
#define EDEADLK 11 /* Resource deadlock avoided */
/* 11 was EAGAIN */
11 is now:
#define EAGAIN 35 /* Resource temporarily unavailable */
Which one that are used, I dont know. And if it differ due to the ABI that
are used.
I believe it is a threading and/or timing issue, since I notice that the
process gets <defunct> on crash at a massive use if 'ps aux | grep hlds'
In the ps manpage you can read:
"Processes marked <defunct> are dead processes (so-called "zombies") that
remain because their parent has not destroyed them properly. These
processes will be destroyed by init(8) if the parent process exits."
Ehm, did I make any sense? Gah, time for a drink with caffeine in it. I am
truly too tired for this...
/Bjorn
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux
---------------------------------------------------------------------------------------
Wanadoo vous informe que cet e-mail a ete controle par l'anti-virus mail.
Aucun virus connu a ce jour par nos services n'a ete detecte.
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlds_linux