[GH-ISSUE #2062] TF2 server segfaults on startup in Fedora 28 x86_64 #1607

New issue

Closed

opened 2026-02-27 02:58:05 +03:00 by kerem · 3 comments

kerem commented

2026-02-27 02:58:05 +03:00

Owner

Originally created by @vectorsigma on GitHub (Oct 29, 2018).
Original GitHub issue: https://github.com/GameServerManagers/LinuxGSM/issues/2062

Describe the bug
In Fedora 28, for Team Fortress 2, replay_srv.so is unable to load, causing segfault and server crash loop in the console. The server will refuse to load.

To Reproduce
Steps to reproduce the behaviour:

download lgsm for tf2 and follow installation instructions in the home directory of a newly created tf2 user.
execute ./tf2server start
execute ./tf2serer console
press enter at the molly guard warning
See the error snippet below:

server_srv.so loaded for "Team Fortress"
Could not load: replay_srv.so
Could not load: replay_srv.so

Looking up breakpad interfaces from steamclient
Calling BreakpadMiniDumpSystemInit
./srcds_run: line 324:  2506 Segmentation fault      (core dumped) $HL_CMD
Add "-debug" to the ./srcds_run command line to generate a debug.log to help with solving this problem
Sat Oct 27 14:03:29 PDT 2018: Server restart in 10 seconds

Expected behaviour
Server software starts without crashing.

Minimum Information
Running the TF2 game, on Fedora 28, x86_64. Installed fresh on 2018/10/27.

Additional context
Turns out, this is just a simple bug because replay_srv.so is compiled to expect the library to be named libcurl-gnutls.so.4, which the Fedora 28 version of libcurl.i686 doesn't provide, by that exact name. You can solve it as follows: ln -nfs /usr/lib/libcurl.so.4 ~/serverfiles/bin/libcurl-gnutls.so.4 as the user you installed the server as.

Proposed fix in #2063.

Originally created by @vectorsigma on GitHub (Oct 29, 2018). Original GitHub issue: https://github.com/GameServerManagers/LinuxGSM/issues/2062 **Describe the bug** In Fedora 28, for Team Fortress 2, `replay_srv.so` is unable to load, causing segfault and server crash loop in the console. The server will refuse to load. **To Reproduce** Steps to reproduce the behaviour: 1. download lgsm for tf2 and follow [installation instructions](https://linuxgsm.com/lgsm/tf2server/) in the home directory of a newly created `tf2` user. 2. execute `./tf2server start` 3. execute `./tf2serer console` 4. press enter at the molly guard warning 5. See the error snippet below: ``` server_srv.so loaded for "Team Fortress" Could not load: replay_srv.so Could not load: replay_srv.so Looking up breakpad interfaces from steamclient Calling BreakpadMiniDumpSystemInit ./srcds_run: line 324: 2506 Segmentation fault (core dumped) $HL_CMD Add "-debug" to the ./srcds_run command line to generate a debug.log to help with solving this problem Sat Oct 27 14:03:29 PDT 2018: Server restart in 10 seconds ``` **Expected behaviour** Server software starts without crashing. **Minimum Information** Running the TF2 game, on Fedora 28, x86_64. Installed fresh on 2018/10/27. **Additional context** Turns out, this is just a simple bug because `replay_srv.so` is compiled to expect the library to be named `libcurl-gnutls.so.4`, which the Fedora 28 version of `libcurl.i686` doesn't provide, by that exact name. You can solve it as follows: `ln -nfs /usr/lib/libcurl.so.4 ~/serverfiles/bin/libcurl-gnutls.so.4` as the user you installed the server as. Proposed fix in #2063.

kerem

2026-02-27 02:58:05 +03:00

closed this issue
added the
type: bug

outcome: issue resolved

distro: CentOS

distro: Fedora
labels

kerem commented

2026-02-27 02:58:06 +03:00

Author

Owner

@tkrn commented on GitHub (Oct 29, 2018):

On RHEL 7.5, I had the same issue.

[tkrn@lanserv:~]$uname -a
Linux lanserv.party.lan 3.10.0-862.14.4.el7.x86_64 #1 SMP Fri Sep 21 09:07:21 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

[tkrn@lanserv:~]$cat /etc/redhat-release
Red Hat Enterprise Linux Server release 7.5 (Maipo)

The following fix resolved the segfault.

ln -s /usr/lib/libcurl.so.4 /home/tf2server/serverfiles/bin/libcurl-gnutls.so.4

After this, it resolved the issue. Please note that on RHEL 7.5 libcurl.so.4 is a default symlink to libcurl.so.4.3.0 which is provided in the 'libcurl' package in the yum repo.

[tkrn@lanserv:~]$ls -lah /usr/lib | grep libcurl
lrwxrwxrwx.  1 root root   16 Jul 30 14:05 libcurl.so.4 -> libcurl.so.4.3.0
-rwxr-xr-x.  1 root root 424K Dec 14  2017 libcurl.so.4.3.0

Just adding a second person confirming the fix.

@tkrn commented on GitHub (Oct 29, 2018): On RHEL 7.5, I had the same issue. ``` [tkrn@lanserv:~]$uname -a Linux lanserv.party.lan 3.10.0-862.14.4.el7.x86_64 #1 SMP Fri Sep 21 09:07:21 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux [tkrn@lanserv:~]$cat /etc/redhat-release Red Hat Enterprise Linux Server release 7.5 (Maipo) ``` The following fix resolved the segfault. `ln -s /usr/lib/libcurl.so.4 /home/tf2server/serverfiles/bin/libcurl-gnutls.so.4` After this, it resolved the issue. Please note that on RHEL 7.5 libcurl.so.4 is a default symlink to libcurl.so.4.3.0 which is provided in the 'libcurl' package in the yum repo. ``` [tkrn@lanserv:~]$ls -lah /usr/lib | grep libcurl lrwxrwxrwx. 1 root root 16 Jul 30 14:05 libcurl.so.4 -> libcurl.so.4.3.0 -rwxr-xr-x. 1 root root 424K Dec 14 2017 libcurl.so.4.3.0 ``` Just adding a second person confirming the fix.

kerem commented

2026-02-27 02:58:06 +03:00

Author

Owner

@dgibbs64 commented on GitHub (Nov 1, 2018):

Issue now merged

@dgibbs64 commented on GitHub (Nov 1, 2018): Issue now merged

kerem commented

2026-02-27 02:58:06 +03:00

Author

Owner

@lock[bot] commented on GitHub (Nov 1, 2019):

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.

@lock[bot] commented on GitHub (Nov 1, 2019): This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.

kerem referenced this issue

2026-02-27 03:54:15 +03:00

[PR #1607] [MERGED] hotfix/170803.1 #3369