Hlds Having A Tanty
Lethy
Join Date: 2002-11-07 Member: 7666Members, Constellation
<div class="IPBDescription">Couldn't get DLL API!</div> Here's the story, my HLDS was running fine with NS, tried my hand at getting ClanMod and AdminMod running on it. After doing so decided to do a clean install of the HLDS (3.1.1.1d) and NS (2.0). I had also just installed the debian <a href='http://packages.debian.org/stable/editors/vim.html' target='_blank'>VIM</a> package which also installed the <a href='http://packages.debian.org/stable/libs/libgpmg1.html' target='_blank'>libgpmg1 package</a> as a dependency.
After the clean install on debian (3.0r1), hlds 3.1.1.1d, ns2.0 (with the replacement ns_i386.so file) I get the following error:
<!--c1--></span><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>CODE</b> </td></tr><tr><td id='CODE'><!--ec1-->retribution:/home/hlds_l# ./hlds_run -game ns +ip 192.168.0.1 +maxplayers 4 +map ns_eclipse -debug
Auto detecting CPU
Using Pentium II Optimised binary.
Enabling debug mode
Auto-restarting the server on crash
Console initialized.
Protocol version 46
Exe version 3.1.1.1
Exe build: 18:20:53 Jul 29 2003 (2469)
WON Auth Server
Server IP address 192.168.0.1:27015
LoadLibrary failed on ./ns/dlls/ns_i386.so: ./ns/dlls/ns_i386.so: invalid ELF header
Host_Error: Couldn't get DLL API from ./ns/dlls/ns_i386.so!
FATAL ERROR (shutting down): Host_Error: Couldn't get DLL API from ./ns/dlls/ns_i386.so!
./hlds_run: line 14: 14319 Segmentation fault (core dumped) $HL_CMD
debug.cmds:1: Error in sourced command file:
Cannot access memory at address 0xbfffcbc0
email debug.log to linux@valvesoftware.com
Sat Aug 9 19:10:39 EST 2003: Server restart in 10 seconds
Sat Aug 9 19:10:41 EST 2003: Server Quit<!--c2--></td></tr></table><span class='postcolor'><!--ec2-->
debug.log is as follows
<!--c1--></span><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>CODE</b> </td></tr><tr><td id='CODE'><!--ec1-->----------------------------------------------
CRASH: Sat Aug 9 19:11:07 EST 2003
Start Line: ./hlds_i686 -game ns +ip 192.168.0.1 +maxplayers 4 +map ns_eclipse -debug
Core was generated by `./hlds_i686 -game ns +ip 192.168.0.1 +maxplayers 4 +map ns_eclipse -debug'.
Program terminated with signal 11, Segmentation fault.
#0 0x00000000 in ?? ()
#0 0x00000000 in ?? ()
End of crash report
----------------------------------------------<!--c2--></td></tr></table><span class='postcolor'><!--ec2-->
I don't believe this is an NS specific problem as it also occurs when trying to run *shudders* CS, of course with the relevant CS .so file in the error message. As I said, this might not be an NS specific problem, but regardless, any help please?
After the clean install on debian (3.0r1), hlds 3.1.1.1d, ns2.0 (with the replacement ns_i386.so file) I get the following error:
<!--c1--></span><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>CODE</b> </td></tr><tr><td id='CODE'><!--ec1-->retribution:/home/hlds_l# ./hlds_run -game ns +ip 192.168.0.1 +maxplayers 4 +map ns_eclipse -debug
Auto detecting CPU
Using Pentium II Optimised binary.
Enabling debug mode
Auto-restarting the server on crash
Console initialized.
Protocol version 46
Exe version 3.1.1.1
Exe build: 18:20:53 Jul 29 2003 (2469)
WON Auth Server
Server IP address 192.168.0.1:27015
LoadLibrary failed on ./ns/dlls/ns_i386.so: ./ns/dlls/ns_i386.so: invalid ELF header
Host_Error: Couldn't get DLL API from ./ns/dlls/ns_i386.so!
FATAL ERROR (shutting down): Host_Error: Couldn't get DLL API from ./ns/dlls/ns_i386.so!
./hlds_run: line 14: 14319 Segmentation fault (core dumped) $HL_CMD
debug.cmds:1: Error in sourced command file:
Cannot access memory at address 0xbfffcbc0
email debug.log to linux@valvesoftware.com
Sat Aug 9 19:10:39 EST 2003: Server restart in 10 seconds
Sat Aug 9 19:10:41 EST 2003: Server Quit<!--c2--></td></tr></table><span class='postcolor'><!--ec2-->
debug.log is as follows
<!--c1--></span><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>CODE</b> </td></tr><tr><td id='CODE'><!--ec1-->----------------------------------------------
CRASH: Sat Aug 9 19:11:07 EST 2003
Start Line: ./hlds_i686 -game ns +ip 192.168.0.1 +maxplayers 4 +map ns_eclipse -debug
Core was generated by `./hlds_i686 -game ns +ip 192.168.0.1 +maxplayers 4 +map ns_eclipse -debug'.
Program terminated with signal 11, Segmentation fault.
#0 0x00000000 in ?? ()
#0 0x00000000 in ?? ()
End of crash report
----------------------------------------------<!--c2--></td></tr></table><span class='postcolor'><!--ec2-->
I don't believe this is an NS specific problem as it also occurs when trying to run *shudders* CS, of course with the relevant CS .so file in the error message. As I said, this might not be an NS specific problem, but regardless, any help please?
Comments
CPU: Dual Pentium-Pro 200MHz (Yeah I know it's not powerful, but it's just a private test server)
RAM: 128MB
OS Version: Debian 3.0r1
Kernel Version: 2.4.18-bf2.4
glibc Version: 2.2.5
HLDS Version: 3.1.1.1d
ns_i386 version: 3.1.1.1 replacement file.
Nop. I'm using the replacement ns_i386.so to run 2.0 on 3.1.1.1 servers from the download page, <a href='http://www.natural-selection.org/ns_i386-3.1.1.1.so.gz' target='_blank'>this link (http://www.natural-selection.org/ns_i386-3.1.1.1.so.gz)</a> to be exact.
This file is corrupt, or there's a serious problem with the dynamic linker. The former is more likely. Did you unzip the NS server zipfile using "unzip -a"? (don't use -a).
*edit* hrm, there's another fellow with a very similar problem. He reports that 3.1.1.1 doesn't crash, but loading up 3.1.1.1d does...
Check the md5sum - mine is:
8ac2133b75e8683c7d8dde1097713f01 ns_i386.so
This file is corrupt, or there's a serious problem with the dynamic linker. The former is more likely. Did you unzip the NS server zipfile using "unzip -a"? (don't use -a).
*edit* hrm, there's another fellow with a very similar problem. He reports that 3.1.1.1 doesn't crash, but loading up 3.1.1.1d does... <!--QuoteEnd--> </td></tr></table><span class='postcolor'> <!--QuoteEEnd-->
I've always made sure I didn't use the -a flag when unzipping the ns ded server.
Meanwhile, I just did a clean install of the HLDS/NS again and well it's working again and I don't believe I have done anything different. I would've thought that maybe the gz file for the replacement .so file is corrupt, but I didn't redownload it and used the same gz file from when it was stuffing up. So my guess is maybe the replacement .so was getting corrupted while being gunzipped and I was just lucky for it to gunzip properly again?
Thanks for everyone's help though.
*edit* hrm, there's another fellow with a very similar problem. He reports that 3.1.1.1 doesn't crash, but loading up 3.1.1.1d does... <!--QuoteEnd--> </td></tr></table><span class='postcolor'> <!--QuoteEEnd-->
Our server is the same way, he updated to 3.1.1.1d on all cs servers, then tried the ns one and it crashed. we are having to run 3.1.1.1 on ns.