[Opacus]

After a few minutes, the server segfaults. I've tried stripping the plugins down to a minimum, but still...

System: Fedora Core 2 Kernel 2.6.5

I quote debug.log here:

---------------------------------------------- CRASH: Mon Nov 15 056:13 CET 2004 Start Line: ./hlds_amd -game cstrike -pingboost 1 +sv_lan 0 +port 27015 +maxplayers 20 +map de_dust2 -secure -debug -pidfile hlds.29793.pid Using host libthread_db library "/lib/tls/libthread_db.so.1". Core was generated by `./hlds_amd -game cstrike -pingboost 1 +sv_lan 0 +port 27015 +maxplayers 20 +map'. Program terminated with signal 11, Segmentation fault. #0 0x5501c126 in ?? () #0 0x5501c126 in ?? () End of crash report ----------------------------------------------

[BAILOPAN]

What Metamod and AMX Mod X version?

[Opacus]

Sorry... Metamod 1.17.3 and AMXX .20 RC7

[BigBaller]

Originally Posted by Opacus Sorry... Metamod 1.17.3 and AMXX .20 RC7

Get 1.0 final and see if that fixes your problems

[BAILOPAN]

also, do you have anything like SELinux/GRSecurity/PAX installed?

[david.vincent]

i'm running metamod 1.7.3 and amxx 1.0 and also seeing alot of seg faults on fedora.

on a fedora fc3 (with selinux) i put in today it was like every 5 minutes when people were playing, if no one was on it was fine. some fedora fc2 servers seem to seg fault a couple of times a day and a suse 9.1 seems to not seg fault.

all running on the same hardware and basically the same counter strike configuration but with half being counter strike and half condition zero. all the stock plugins and modules loaded. when the seg faults started (after changing to fedora with rc releases od amxx) i played around with disabling all the plugins/modules but it didn't seem to make a difference).

i've changed back to adminmod on the server that was crashing every 5 minutes and it's been up for 2 hours with lots of players.

heres the log of a crash:

L 11/17/2004 - 20:02:03: "Lagomorph<8><STEAM_0:1:419566><CT>" triggered "latency" (ping "37")
L 11/17/2004 - 20:02:03: "Lagomorph<8><STEAM_0:1:419566><CT>" disconnected
Dropped Lagomorph from server
Reason: Client sent 'drop'
Razze| : /rank
L 11/17/2004 - 20:02:18: "Razze|<12><STEAM_0:0547126><TERRORIST> " say "/rank"
L 11/17/2004 - 20:02:22: "d0j00<6><STEAM_0:0:1983482><TERRORIST>" killed "AFK!<4><STEAM_0:0:546646><CT>" with "m4a1"
L 11/17/2004 - 20:02:22: Team "TERRORIST" triggered "Terrorists_Win" (CT "4") (T "6")
L 11/17/2004 - 20:02:22: World triggered "Round_End"
Razze| : 2
L 11/17/2004 - 20:02:24: "Razze|<12><STEAM_0:0547126><TERRORIST> " say "2"
./hlds_run: line 341: 27725 Segmentation fault (core dumped) $HL_CMD
debug.cmds:1: Error in sourced command file:
Cannot access memory at address 0x368f39c
email debug.log to [email protected]
Wed Nov 17 20:02:27 NZDT 2004: Server restart in 10 seconds
Updating server using Steam.
Checking bootstrapper version ...
Updating Installation
Logging in as 'xxx'
Checking/Installing 'Counter-Strike Base Content' version 15

CRASH: Wed Nov 17 20:02:27 NZDT 2004
Start Line: ./hlds_i686 +port 27015 -pingboost 1 -steamuser xxx -steampass xxx -debug -autoupdate -retry -game cstrike +exec cstrike.cfg -pidfile hlds.27667.pid
Using host libthread_db library "/lib/tls/libthread_db.so.1".
Core was generated by `./hlds_i686 +port 27015 -pingboost 1 -steamuser xxx -steampass xxx'.
Program terminated with signal 11, Segmentation fault.
#0 0x0060b08d in ?? ()
#0 0x0060b08d in ?? ()
End of crash report

[M.C.S.]

I had similar problems with SuSE Linux 9.0 (crash after some minutes). On other machines I never had any problems

[BAILOPAN]

It seems the JIT still has problems on Fedora. I will release a special binary for you guys with the JIT disabled and hopefully that will work.

[M.C.S.]

Seen on a Debian testing machine:

----------------------------------------------
CRASH: Wed Nov 17 12:41:45 CET 2004
Start Line: ./hlds_amd -console -debug -game cstrike -pingboost 1 -tos +ip 217.1
72.187.195 +port 27015 +map de_dust +maxplayers 18 -pidfile hlds.504.pid
Using host libthread_db library "/lib/tls/libthread_db.so.1".
Core was generated by `./hlds_amd -console -debug -game cstrike -pingboost 1 -to
s +ip 217.172.187.195'.
Program terminated with signal 11, Segmentation fault.
#0 0x4002a87f in ?? ()
#0 0x4002a87f in ?? ()
End of crash report
----------------------------------------------

First this has been a SuSE 9.0 server, but because of these messages I switched to Debian Sarge. But same problem there. I get the same messages with AMXX 0.16 so I don't think the JIT causes the problems. Maybe a hardware specific problem? Or a special shared object? I'd like to test a version where all needed libraries are compiled into the .so if possible

[BAILOPAN]

It's fishy that all the systems this happens on are using the NPTL (/lib/tls) libraries.

Compiling in all the necessary libraries is a bad idea... I will post some test binaries here later today though

Better yet, anyone feel like giving me shell access to run GDB and track this down?