djh, Could you type "ulimit -c unlimited" before starting the server and then try it again. This sets the max size for core dumps to unlimited instead of 2MB, so it will contain the contents of all of the memory used by hlds.
Once you get that, post the debug.log here and a link to the core dump. -- Andy On Thu, 26 Mar 2009 19:00:59 -0700, "Tony Paloma" <[email protected]> said: > If you want to put all the arguments to the program being debugged on the > command line like that you need to use the gdb switch --args. > > Try: > gdb --args /home/services/hlds/ab_test/hlds_amd -game cstrike +map > de_dust2 +ip 192.168.20.37 +port 27035 > > -----Original Message----- > From: [email protected] > [mailto:[email protected]] On Behalf Of djh > Sent: Thursday, March 26, 2009 4:36 PM > To: Half-Life dedicated Linux server mailing list > Subject: Re: [hlds_linux] random segfault issue > > > Hello > Selinux is disabled > i tried running gdb as you said but i cant get it to work. > I get: > gdb /home/services/hlds/ab_test/hlds_amd -game cstrike +map de_dust2 +ip > 192.168.20.37 +port 27035 > gdb: unrecognized option `-game' > Use `gdb --help' for a complete list of options. > > i noticed hlds_run has a -debug option and by default it should use gdb > so i > added that to the start command and waited > it produced the following log and a coredump: > cat debug.log > ---------------------------------------------- > CRASH: Fri Mar 27 01:16:46 EET 2009 > Start Line: ./hlds_amd -debug -game cstrike -timeout 2 -pidfile ab_kz.pid > +map kz_kzdk_templebhop +ip 192.168.20.41 +maxplayers 20 -pingboost 3 > +sys_ticrate 1000 +sv_region 3 +sv_enableoldqueries 1 -pidfile ab_kz.pid > Using host libthread_db library "/lib/libthread_db.so.1". > Core was generated by `./hlds_amd -debug -game cstrike -timeout 2 > -pidfile > ab_kz.pid +map kz_kzdk_temp'. > Program terminated with signal 11, Segmentation fault. > #0 0x007b7560 in ?? () > #0 0x007b7560 in ?? () > No symbol table info available. > No shared libraries loaded at this time. > Stack level 0, frame at 0xbff5db64: > eip = 0x7b7560; saved eip End of crash report > ---------------------------------------------- > > the coredump is at http://79.134.32.23/cs/core.26960.tgz > > thank you for your reply :) > > > > ----- Original Message ---- > From: "[email protected]" > <[email protected]> > To: Half-Life dedicated Linux server mailing list > <[email protected]> > Sent: Thursday, March 26, 2009 9:47:14 PM > Subject: Re: [hlds_linux] random segfault issue > > Error 4 is "Interrupted system call". (See > /usr/include/asm-generic/errno-base.h for the messages that match the > error numbers.) > > Do you have SELinux enabled by chance? > > > As for debugging, try running it with gdb and post the output back to > the list: > > # gdb /path/to/hlds_amd +map blah > (gdb) set follow-fork-mode child > (gdb) run > (wait for it to crash) > (gdb) bt > > > -- > Andy > > > On Wed, 25 Mar 2009 14:04:30 -0700 (PDT), "djh" <[email protected]> > said: > > > > Hi. > > I have been experiencing random segfaults on my server, > > i run 4 hlds (cs) and 3 srcds (css) > > > > machine specs are as follows: > > > > Sun Fire X4200 M2 server > > OS: CentOS 5.2 32bit > > 2x Dual-Core AMD Opteron(tm) Processor 2216 @ 2393.638 Mhz > > MemTotal: 4151164 kB > > Linux cs2pps 2.6.25.10-c-p-v0.6 #13 SMP Tue Mar 17 17:56:53 EET 2009 i686 > > athlon i386 GNU/Linux > > > > > > > > servers fail apparently at random (i checked logs, couldnt find anything) > > the only thing i have is the message in /var/log/messages > > > > Mar 25 18:20:30 cs2pps kernel: hlds_amd[29289]: segfault at 4 ip b7abf560 > > sp bfd7a550 error 4 in engine_amd.so[b7a35000+ba000] > > > > and similar to that > > please at least tell me how to debug it :) - for example what does > > error 4 stand for ? > > thank you :) _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds_linux
