Gorch 0 Posted June 16, 2003 As the topic indicates, the OFP dedicated server (1.91c) first ran on a Suse 8.0 linux system. It worked fine, no problems at all. After upgrading the linux operating system (Suse 8.0 -> 8.2) the problems started. When trying to start the dedicated server from the console i get 17:28:00 Dedicated server created, memory used: 4296 KB Segmentation fault Installing the OFP server from scratch doesnt help, i get the same error. I also tried to install the OFP server on a diffrent machine running Suse 8.2, with the same result. After reading several topics on 'linux server problems', i tried strace. This are the last lines of output: open("/lib/libc.so.6", O_RDONLY) = 38 read(38, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0pY\1\000"..., 1024) = 1024 fstat64(38, {st_mode=S_IFREG|0755, st_size=1491599, ...}) = 0 old_mmap(NULL, 1268004, PROT_READ|PROT_EXEC, MAP_PRIVATE, 38, 0) = 0x580dd000 mprotect(0x5820c000, 26916, PROT_NONE) = 0 old_mmap(0x5820c000, 16384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 38, 0x12f000) = 0x5820c000 old_mmap(0x58210000, 10532, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x58210000 close(38) = 0 open("/lib/ld-linux.so.2", O_RDONLY) = 38 read(38, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\200\n\0"..., 1024) = 1024 fstat64(38, {st_mode=S_IFREG|0755, st_size=97518, ...}) = 0 old_mmap(NULL, 83576, PROT_READ|PROT_EXEC, MAP_PRIVATE, 38, 0) = 0x58213000 mprotect(0x58227000, 1656, PROT_NONE) = 0 old_mmap(0x58227000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 38, 0x14000) = 0x58227000 close(38) = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++ I cant figure out what the problem is with this... Someone noted there may be prolems with some versions of gcc (compiler? library?). Suse 8.2 is using gcc version 3.3 (20030226 prerelease version). Could this be the problem? Is there any way to give the OFP server some other, working version of gcc without having to resetup the whole linux system? Share this post Link to post Share on other sites
benu 1 Posted June 16, 2003 The compiler problems were only affecting the compilation of the tolower tool while installing the server, not running it. Do you know of any changes Suse might have made? Maybe some of those kernel security patches? I have read through the documentation of some of those patches and there are patches that break functionality for example <table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE"> XFree86 still will not operate correctly with this option enabled, so DO NOT CHOOSE Y IF YOU USE XFree86. or <table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE"> Enabling this feature could make certain apps like VMWare stop working, as they need to write to other locations in /dev/mem. or (Non-executable Mem-Pages from GR-Security-Patch, seems to be the most likely to me, but i am not really an expert) <table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE"> This will also break programs that rely on the old behaviour and expect that dynamically allocated memory via the malloc() family of functions is executable (which it is not). Notable examples are the XFree86 4.x server, the java runtime and wine. Try getting an unpatched kernel from ftp.kernel.org and using that... or switch to debian or (even better) gentoo ;) Share this post Link to post Share on other sites
Adammo 0 Posted September 14, 2003 Here is my log on the RH version 9 linux server: 18:32:45 Roles assigned 18:32:45 Reading mission ... 18:33:05 Mission read 18:33:06 Admin [iCM]-Moore logged in. 18:34:07 Game started 18:48:29 Player [iCM]-Moore disconnected 18:48:29 Admin [iCM]-Moore logged out. 18:51:55 Player Cpt Moore connecting 18:52:01 Player Cpt Moore connected (id=195) Segmentation fault [root@ ofp]# It appears they we are also getting this same error on Red HAT 9. -Adammo Share this post Link to post Share on other sites
vesuvio 0 Posted September 19, 2003 For Segmentation Fault user Is know that the original glibc version 2.3.2 have a bit problems with some applications. Some new released distro (redhat,suse,....) use the libc customize and this can give some problem with server ofp. In the our case first of update distro look for version glibc installed if >2.3.2 the server start well This problem is possible resolve only with optimize the source code server ofp for new glibc. REDHAT 7.3 ---> Ok REDHAT 8.0 ---> Ok REDHAT 9.x ---> Failed SUSE 8.0 -------> Ok SUSE 8.2 -------> Failed Getoo -----------> Ok Trustix 2.0 -----> Failed Trustix 1.5 ------> OK Vesuvio [MCY] Share this post Link to post Share on other sites
shock604 0 Posted October 4, 2003 Vesuvio, are you trying to say that if the version of glibc is LESS than 2.3.2 then the server will work? 2.3.2 is the highest version available so even though your post suggests it. You can't have a higher version than 2.3.2. I'm getting segmentation faults on my linux box with 2.3.1 btw. I think it only works with version 2.3 =\ Share this post Link to post Share on other sites
benu 1 Posted October 4, 2003 No, i think he is trying to say that some distros use modified glibc. Gentoo uses 2.3.2 too and it works there. I seem to recall something about some newer redhat and/or suse distros using threads by default, maybe it's got something to do with that, but i am just guessing here... Share this post Link to post Share on other sites
shock604 0 Posted October 5, 2003 Ah, I see. Well Mandrake 9.0 doesn't seem to like the server binary either. I'm going to try Debian 3.0 see if that works. Share this post Link to post Share on other sites
vesuvio 0 Posted October 6, 2003 The segmentation fault on some Linux system is possible resolve only with optimizing the source code of the server ofp for that glibc. See Halflife linux server Quote[/b] ] » Fixed DMC powerup bug where the invisibility ring would make people glow red.» Fixed rcon bug in HLTV. » Fixed delay being reset to default value during changelevel. » Fixed a problem with propagating HLTV banners. » Fixed client freeze on exit. » Fixed framerate problem on some ATI video cards. » Fixed MaxPing filter not being saved correctly. » Fixed sky box bug. » Fixed TFC spy disguise menu bug. » Fixed DMC view model animation bug. » Fixed bug using "exec " on large files (e.g. banned.cfg). » Linux: Fixed server using 64 MB more memory than it needed to. » Linux: Fixed 'undefined symbol: __strtouq_internal' on some Linux systems. » Linux: Fixed segmentation fault on some Linux systems Share this post Link to post Share on other sites
jmckind 0 Posted October 13, 2003 I have setup a Windows box for my OFP server until I can get this figured out or BIS releases a server update. With OFP2 coming out on the horizon though, I doubt BIS will do much more for the OFP server Share this post Link to post Share on other sites
benu 1 Posted October 14, 2003 I have setup a Windows box for my OFP server until I can get this figured out or BIS releases a server update. With OFP2 coming out on the horizon though, I doubt BIS will do much more for the OFP server There just was a new ofp dedicated (1.94) one day ago. And quite frankly, as much as i love linux, if you use a linux brand that is binary incompatible to the version before you have to live with things like that. There are lots of linux brands that still work, so it's more the fault of the suse/redhat version you use than of bis. I use gentoo, which always installs the newest versions of everything available and ofp works for me. If suse/redhat use some modified versions which are not compatible (anyone remember redhat using an inofficial gcc version for some versions of their linux?) and you are unhappy about it then switch to a vendor which is compatible... Share this post Link to post Share on other sites
vesuvio 0 Posted October 14, 2003 yeah benu  I agree with you. The MCY have now 1.94 Linux Server for test Vesuvio [McY] Share this post Link to post Share on other sites
Adammo 0 Posted October 14, 2003 Here is a thought.. I run redhat 9 and get the error I run redhat 8.0 and get the error.. I don't beleive I have gotten the error after a reboot and then start the server... seems clean... if I stop the server and restart it.. this seems to increase my chance of getting the error.. I sent a strace to BI to look at while I was getting the error and found that the server thought that these ports were still in use. Maybe the process is not shutting down or timing out prompt enough.. so when then serve is restarted the server crashes.. just a thought. Share this post Link to post Share on other sites
benu 1 Posted October 15, 2003 My server (hardware) gets restarted once in a month at max. I have ni problems restarting the ofp server. I have a 5 second pause in the restart script between the killing of the old processes and the starting of the new ones. Share this post Link to post Share on other sites
shortylaforge 0 Posted October 20, 2003 As the topic indicates, the OFP dedicated server (1.91c) first ran on a Suse 8.0 linux system. It worked fine, no problems at all. After upgrading the linux operating system (Suse 8.0 -> 8.2) the problems started. When trying to start the dedicated server from the console i get 17:28:00 Dedicated server created, memory used: 4296 KB Segmentation fault Installing the OFP server from scratch doesnt help, i get the same error. I also tried to install the OFP server on a diffrent machine running Suse 8.2, with the same result. After reading several topics on 'linux server problems', i tried strace. This are the last lines of output: open("/lib/libc.so.6", O_RDONLY)     = 38 read(38, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0pY\1\000"..., 1024) = 1024 fstat64(38, {st_mode=S_IFREG|0755, st_size=1491599, ...}) = 0 old_mmap(NULL, 1268004, PROT_READ|PROT_EXEC, MAP_PRIVATE, 38, 0) = 0x580dd000 mprotect(0x5820c000, 26916, PROT_NONE)  = 0 old_mmap(0x5820c000, 16384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 38, 0x12f000) = 0x5820c000 old_mmap(0x58210000, 10532, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x58210000 close(38)                = 0 open("/lib/ld-linux.so.2", O_RDONLY)   = 38 read(38, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0\200\n\0"..., 1024) = 1024 fstat64(38, {st_mode=S_IFREG|0755, st_size=97518, ...}) = 0 old_mmap(NULL, 83576, PROT_READ|PROT_EXEC, MAP_PRIVATE, 38, 0) = 0x58213000 mprotect(0x58227000, 1656, PROT_NONE)  = 0 old_mmap(0x58227000, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED, 38, 0x14000) = 0x58227000 close(38)                = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++ I cant figure out what the problem is with this... Someone noted there may be prolems with some versions of gcc (compiler? library?). Suse 8.2 is using gcc version 3.3 (20030226 prerelease version). Could this be the problem? Is there any way to give the OFP server some other, working version of gcc without having to resetup the whole linux system? Problem: Linux Dedicated OFP Server under SuSE8.2 cannot be created. Error: Segmentation fault Solution: <table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE">Define dedicated OFP User incl. Home Environment Useradd ofpadmin -m <table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE">Inst. OFPLinux Server like default into $HOME/ofplinux <table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE">Run unshar server-xxx.shar to install and LowerCase the Directory <table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE">Mkdir $HOME/ofplinux/lib <table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE">Copy the following Libs from Suse8.0  /lib to $HOME/ofplinux/lib, and change Ownership to ofpadmin:users ld-linux.so.2 libc.so.6 libnss_dns.so.2 libnss_files.so.2 libresolv.so.2 Don't copy them into /lib !!! Now u must edit the following section in the OFPSERVER Script: Code::: ++++++++++++++++++++++++++++++++++++++++++++++   cd ${OFP_DIR}   echo >>${LOGFILE} "WATCHDOG ($$): [$(date)] Starting server (port ${PORT})..."   (export LD_LIBRARY_PATH=/home/ofpadmin/ofplinux/lib;${SERVER} >>${LOGFILE} 2>&1 -server -config=${CONFIG} -port=${PORT} -pid=${PIDFILE})   if [ -f ${RUNFILE} ]; then echo >>${LOGFILE} "WATCHDOG ($$): [$(date)] Server died, waiting to restart..." unset LD_LIBRARY_PATH sleep 5s   else echo >>${LOGFILE} "WATCHDOG ($$): [$(date)] Server shutdown intentional, watchdog terminating" unset LD_LIBRARY_PATH   fi done   ;; ++++++++++++++++++++++++++++++++++++++++++++++ Run the Server with user: Ofpadmin And Startscript:<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE"> ofpserver start DON'T DO THIS WITH UR ROOT ACCOUNT!!!!! PS: I upload the Source Files this evening onto my WebSpace. This 'workaround' is equivalent to other Linux Distributions. Ive tested the Server 8h, and it works fine *gg*. @BIS: plz. upgrade the SERVER binary to libc 2.3.2 Thats it folks!! Regards, Shortylaforge OFP rulez!!! Share this post Link to post Share on other sites
benu 1 Posted October 20, 2003 @BIS: plz. upgrade the SERVER binary to libc 2.3.2 OFP works fine on my gentoo using libc 2.3.2. Suse just ships a broken version. Share this post Link to post Share on other sites
shortylaforge 0 Posted October 20, 2003 But it isn't an only SuSE Problem. Ive heared about RedHat 9 ect... Its only a workaround for guys which can't change the Linux Distribution like me. Share this post Link to post Share on other sites
benu 1 Posted October 20, 2003 Afaik ONLY with new suse and redhat versions. And those two are well-known for modifying core components like kernel, gcc, etc... Share this post Link to post Share on other sites
vesuvio 0 Posted October 21, 2003 @ Suma Please report at developer staff server linux of optimize server for glibc 2.3.2. On Some distro linux not work the server. Look on debug libc.so.6. Thx Now all the know new linux distro use glibc 2.3.2. Imao benu Who have do the tests writed on ? Vesuvio [MCY] Share this post Link to post Share on other sites
benu 1 Posted October 21, 2003 Gentoo uses glibc 2.3.2. OFP works on gentoo. -> OFP works with glibc 2.3.2. Suse and Redhat modify the kernel/gcc/libc. OFP does not work with Suse and Redhat. -> Use a linux distri that is not broken. Share this post Link to post Share on other sites
H-Milch1 0 Posted November 23, 2003 No other solution for this problem isp. SUSE 9? Share this post Link to post Share on other sites
vesuvio 0 Posted November 24, 2003 You try NSCD demon on. Now On our server run redhat9.0 we have first update with new glibc package relesead of redhat, (resolve much problems) and after abilitated nscd demon. Remember Nscd is a daemon that provides a cacheing of glibc and for the passwd, group, and hosts databases through standard libc interfaces, such as getpwnam, getpwuid, getgrnam, getgrgid, gethostbyname, and others. Each cache has a separate TTL (time-to-live) for its data; modifying the local database (/etc/passwd, and so forth) causes the cache to become invalidated within fifteen seconds. Note that the shadow file is specifically not cached. getspnam calls remain uncached as a result. Vesuvio [McY] Share this post Link to post Share on other sites
benu 1 Posted December 4, 2003 I just had time (and especially space on my harddrive) to test on gentoo again, and i have to retract my previous statement: ofp does not work there anymore (using glibc -2.3.2-r3). This does not really affect me, as GDT server is running on debian, but for the record... BTW: i do not use nscd (never did, even when the server was still running with redhat) and it always worked... Share this post Link to post Share on other sites
darkpeace 0 Posted October 9, 2004 Yeah I am having issues aswell, even using; ENV LD_ASSUME_KERNEL=2.4.1; (See first few chapters of SuSe 9.1 Professional Admin Guide for information, maybe try Google aswell on LD_ASSUME_KERNEL) Get Segmentation Fault when not using -nomap (and it tries to use around 256mb when loading) Get Sockets error when using -nomap (and it uses exactly 4096 kb when loading) Server name has no spaces or hyphens, just; twilightofp I heavily doubt the DNS / hostname is incorrectly setup, is there anything anyone knows of that might be causing the issue though ? Tried with Firewall off aswell. Might be due do newer Kernels reserving memory (or LD_ASSUME_KERNEL is doing jack all for ofpserver), and only letting it use 4096 kb (without -nomap) and this thus leads to problems. I am getting to the point where I am thinking of installing a 2.4.1 Kernel under a http://www.vmware.com virtual machine ? Any confirmed 100% OFP 1.96 working Linux Distros we know of ? (do I need to dig up Red Hat 7.2 for example) ? I read that someone in these forums was getting 55fps using SuSe 9.0 (or 9.1) Professional... might be a load of bull, maybe they know something we don't. Share this post Link to post Share on other sites
killswitch 19 Posted October 9, 2004 Using Fedora Core 2/x86_64 here on a 2.6.x kernel with a LD_LIBRARY_PATH "trick" in conjunction with a set of older libs. That trick is described here: "1.96 on Suse" And yes, it would be very nice to have the linux server be recompiled with more up to date compilers and libs. Also, to get the 1.96a Win32 fixes in there. Share this post Link to post Share on other sites
benu 1 Posted October 9, 2004 I am using gentoo with kernel 2.6.7 (2.6.8 after next server reboot) and LD_LIBRARY_PATH. No problems here. Server uses 111/122mb of memory. The kernels from redhat and suse are heavily modified with inofficial patches, try a vanilla from ftp.kernel.org. But afaik suse has modified the rest of the distro also, so you may get troubles when using an unmodified linux kernel. You can argue if you want to call suse still linux Share this post Link to post Share on other sites
