Segfaults everywhere

[jheiv]

I am getting segfaults once every 6-8 hours.

Some things to note:
The problem isn't limited to two installations: two installations on the same box have the same kind of faults.

The problem isn't limited to Intel or AMD: I get the segfaults on my private dedicated server (specs below) which runs an Intel processor as well as the same segfaults on my local box which runs an AMD processor.

The problem isn't due to a plug-in, I get segfaults on both servers running mani-admin and completely fresh vanilla installs with no plugins.

A commonality is that both boxes run Fedora 11, although I should mention that the Intel box used to run Fedora 8 and had these segfaults too.

We've had segfaults like:
Segfault: server_i486.so:
-- [01] NavAreaBuildPath<ShortestPathCost>
-- [03] NavAreaBuildPath<HostagePathCost>
-- [08] CUtlRBTree<CUtlMap<char const*, int, int>::Node_t, int, CUtlMap<char const*, int, int>::CKeyLess>::NextInorder ()
-- [11] CBaseEntityOutput::~CBaseEntityOutput$base ()
-- [14] NavAreaBuildPath<ShortestPathCost> ()
-- [17] CUtlLinkedList<CNavArea*, int>::RemoveAll ()
-- [23] NavAreaBuildPath<HostagePathCost> ()
-- [24] NavAreaBuildPath<HostagePathCost> ()
-- [25] CUtlLinkedList<SpotOrder, int>::RemoveAll ()
-- [33] NavAreaBuildPath<HostagePathCost>(CNavArea*, CNavArea*, Vector const*, HostagePathCost&, CNavArea**) ()
-- [35] NavAreaBuildPath<HostagePathCost>(CNavArea*, CNavArea*, Vector const*, HostagePathCost&, CNavArea**) ()



Segfault: vphysics_i486.so
-- [02] IVP_Real_Object::update_exact_mindist_events_of_object ()
-- [04] IVP_OO_Watcher::~IVP_OO_Watcher$delete ()
-- [05] IVP_Mindist::mindist_hull_limit_exceeded_event ()
-- [06] IVP_Real_Object::update_exact_mindist_events_of_object ()
-- [07] IVP_Real_Object::update_exact_mindist_events_of_object ()
-- [10] IVP_Mindist_Manager::recheck_ov_element ()
-- [12] IVP_Mindist::mindist_hull_limit_exceeded_event ()
-- [13] IVP_Real_Object::update_exact_mindist_events_of_object ()
-- [15] IVP_Mindist::~IVP_Mindist$base ()
-- [18] IVP_Mindist_Manager::recheck_ov_element ()
-- [19] IVP_Mindist::mindist_hull_limit_exceeded_event ()
-- [20] IVP_Mindist::mindist_hull_limit_exceeded_event ()
-- [21] IVP_Hull_Manager::delete_hull_manager ()
-- [26] IVP_Mindist::mindist_hull_limit_exceeded_event(float) ()
-- [27] IVP_Mindist::mindist_hull_limit_exceeded_event(float)[ ()
-- [28] IVP_Mindist_Base::~IVP_Mindist_Base() ()
-- [30] IVP_OV_Element::remove_oo_collision(IVP_Collision*)
-- [32] IVP_Mindist::mindist_hull_limit_exceeded_event(float) ()
-- [34] IVP_Mindist::mindist_hull_limit_exceeded_event(float) ()

Where the number in square brackets is just a counter to reference the crash dump.

Any Ideas??

Please help!

Thanks to anyone who has any idea.


Private Dedicated Server:
Server OS: Fedora 11, 32-bit PAE Kernel
Processor: Intel® Xeon® CPU X3320 @ 2.50GHz
Ram: 2GB
Game(s): Source Dedicated Server (srcds)
Start Up Command: ./srcds_run -autoupdate -game cstrike -tickrate 66 -ip 66.135.40.174 -port 27015 -verify_all +map de_aztec -debug"
Admin Mods: I have tried both a vanilla install with no mods as well as an install with just mani admin -- both experience this problem.

1. Show your command line If it crashes upon start up when show that as well.
It does not crash on startup and when it crashes it immediately restarts. The game logs show that it was in the middle of calculating stats (weapon_stats, weapon_stats2, etc). The crashes dump crash cores, which I have run through GDB and attached three of.

2. Please include debug.log (to get debug.log you need gdb installed and 0debug in your command line)
N/A - debug.log doesn't have anything other than the time of the crash and the command line used -- see the attached dump details for more details.

3.What distro are you running?
Fedora 11 32-bit

4.GCC Version
$ gcc --version
gcc (GCC) 4.4.1 20090725 (Red Hat 4.4.1-2)

5.glibc version
$ ldd --version
ldd (GNU libc) 2.10.1
Copyright © 2009 Free Software Foundation, Inc.

6. Kernel Version
$ uname -r
2.6.30.8-64.fc11.i686.PAE

Gameserver Versions
$ ./steam -command version
Bootstrapper version: 34
Versions installed in '.':
Source Dedicated Server Linux: 96
Base Source Shared Materials: 8
Base Source Shared Models: 4
Base Source Shared Sounds: 4
Counter-Strike Source Shared Content: 74

All glibc packages
$ rpm -qa | grep glibc
glibc-debuginfo-2.10.1-5.i686
glibc-2.10.1-5.i686
glibc-common-2.10.1-5.i586
glibc-headers-2.10.1-5.i586
glibc-devel-2.10.1-5.i586
glibc-debuginfo-common-2.10.1-5.i586

Full GCC Version
$ gcc -v
Using built-in specs.
Target: i586-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-bootstrap --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-languages=c,c++,objc,obj-c++,java,fortran,ada --enable-java-awt=gtk --disable-dssi --enable-plugin --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-1.5.0.0/jre --enable-libgcj-multifile --enable-java-maintainer-mode --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --disable-libjava-multilib --with-ppl --with-cloog --with-tune=generic --with-arch=i586 --build=i586-redhat-linux
Thread model: posix
gcc version 4.4.1 20090725 (Red Hat 4.4.1-2) (GCC)

[jheiv]

Some more questions:

1) The permissions on my shared libraries (server_i486.so, vphysics_i486.so, etc) are all:
-rw-r--r-x

Do those permissions even make sense? Why would the owner not be able to execute it, but "other" be able to execute? I've read that most times you don't need the DSOs to be executable, but would making them executable help?

2) I have segfaults from server_i486.so, but the process running is srcds_i686 -- so shouldn't it load the server_i686.so library?

3) Fedora users: has anyone gotten it to work on F11 without anything non-stock -- like kernel modifications, etc?

4) I see that srcds_run adds the "bin" directory to the LD_LIBRARY_PATH -- anyone ever try adding the cstrike/bin directory (containing server_i486.so)?

Any help / guidance on next debugging steps / etc is greatly appreciated.

Thanks in advance!

[BehaartesEtwas]

I had similar segfaults (the first one) at mapchanges, until I regenerated my one nav-files (they are also used for the hostages, so you need them even if you don't use bots on your server). try deleting (or better moving) all .nav files from your maps directory, then start every map on your server with bots. that will generate new .nav files. unfortunately bomb spots might be swapped on the radar after this, but for public servers this might be ok.