Segmentation fault core dumped.
- Thread starter Claudiim
- Start date
Check if this problem is fixed:
C++ - Crash Party TFS 0.4 (8.60)
If you run it on linux, first install 'gdb' (debugger):
PHP:
sudo apt install gdb
PHP:
ulimit -c unlimited
PHP:
./otxAfter next crash there should appear file named 'core' in folder with binary file. Then you can use it to detect problem.
I can try to help you, add me on Skype: [email protected]
marcelomnzs
Member
- Joined
- Aug 9, 2019
- Messages
- 121
- Reaction score
- 17
I did the above precedent but I can't find the log file after another core dump. where is it please
marcelomnzs
Member
- Joined
- Aug 9, 2019
- Messages
- 121
- Reaction score
- 17
i got it, ubuntu 22. but i think is ok right?:
apport.service is not a native service, redirecting to systemd-sysv-install.
Executing: /lib/systemd/systemd-sysv-install disable apport
pink_panther
Excellent OT User
if you're not running apport, it should save a core if it crashes again if you're using gdb.
Can also check current directory, or /var/crashes for core dumps.
but you absolutely need a core dump and backtrace log, the error message only is completely useless unfortunately.
It's fine. It means it was enabled and now it's disabled. It should now save
core file in folder where you run application.You can check where it will store
core file by executing:
Code:
cat /proc/sys/kernel/core_patterncore, it means it will save in folder where you run app.
marcelomnzs
Member
- Joined
- Aug 9, 2019
- Messages
- 121
- Reaction score
- 17
thanksss!!!
edit: (no core dumped this time)
shit.. I was looking at the console now, and this morning I had this problem, similar to coredump... yesterday I did another update, but I completely closed restart.sh and opened it again. What can it be?
my restart.sh:
other edit:
i got it:
but in the folder with the tfs have nothing by gdb, like "core"
edit3:
in /var/crash/ have a log:
.. help pls? :s
edit4: the timestamp of this log is by 2days ago... :s
Last edited:
@marcelomnzs
Report in
Looks like your restarter is missing very important setting:
by default core file can be generated only for applications that use less than 8MB RAM. OTS uses gigabytes, so it cannot generate
By running command above you tell linux that for code executed after that command [start of OTS] size of core file is unlimited.
Your restarter should be:
You got to restart OTS AND restarter itself [by killing it using CTRL+C and starting again], to make that change in
Then you can test, if your OTS generates
First find OTS process ID:
(replace
It will return something like:
second column is process ID, in my case it's
Then kill that process with signal 11
It will stop OTS with kind of error as when server crash.
If your server generates
Report in
/var/crash/ is probably from old crashes that happened before you disabled apport.Looks like your restarter is missing very important setting:
Code:
ulimit -c unlimitedcore file.By running command above you tell linux that for code executed after that command [start of OTS] size of core file is unlimited.
Your restarter should be:
Code:
ulimit -c unlimited
while true;
do ./tfs;
donerestart.sh work, not only restart OTS.Then you can test, if your OTS generates
core file on crash by crashing it using kill command.First find OTS process ID:
Code:
ps -aux | grep tfstfs with name of file you use to run OTS, it can be otservbr, theforgottenserver etc.)It will return something like:
Code:
root@arm:~# ps -aux | grep tfs
root 105 0.0 0.0 0 0 ? S May09 0:00 [ecryptfs-kthrea]
root 1011530 6.1 1.2 461468 301632 pts/1 Sl+ Oct18 3872:01 ./tfs
root 1912976 0.0 0.0 7700 676 pts/0 R+ 22:21 0:00 grep --color=auto tfs1011530.Then kill that process with signal 11
Code:
kill -11 1011530If your server generates
core file, it should generate it.
Last edited:
marcelomnzs
Member
- Joined
- Aug 9, 2019
- Messages
- 121
- Reaction score
- 17
just informing. running this command asked for px installation
apt install px
running it again it doesn't tell me anything.
through htop I got the PID and executed kill -11 PID
and had a core dump, and a log generated.
now I will wait for the next crash to analyze. thank you very much @Gesior.pl
apt install px
running it again it doesn't tell me anything.
through htop I got the PID and executed kill -11 PID
and had a core dump, and a log generated.
now I will wait for the next crash to analyze. thank you very much @Gesior.pl
marcelomnzs
Member
- Joined
- Aug 9, 2019
- Messages
- 121
- Reaction score
- 17
nice man.. i'm looking the core."pid" but i cant open.. how read the error to analyze after??Updated code above. It'sps, notpx
edit:
other thing. in htop I notice have 4 process to tfs..
the first one is the same appear on "ps -aux | grep tfs" and have 3 others...
i have only 1 cpu in this vps. is it right?
i have only 1 cpu in this vps. is it right? - yes, but TFS 1.x runs on 4 threads (network thread, dispatcher thread, scheduler thread, database async thread), which are reported as processes on LinuxYour CPU can run any number of processes/threads, CPU will run each for 0.000X sec, stop it, run other and loop over them to make it look like they all run in same time.
how read the error to analyze after?? - gdb core tfs (or gdb tfs core I don't remember order) and then in gdb terminal bt full to generate report
marcelomnzs
Member
- Joined
- Aug 9, 2019
- Messages
- 121
- Reaction score
- 17
ok, i will check when cracsh it..
relating to the other topic, do you think it is possible that what is causing a core dump is some attack? I'm going to start applying the rules..
relating to the other topic, do you think it is possible that what is causing a core dump is some attack? I'm going to start applying the rules..
Only thing that can cause crash is bug in C++ (engine) or Lua (on TFS 1.x where Lua bugs can crash engine).
Of course, some players know engine bugs and abuse them to crash OTSes, to clone items.
In some edge scenario: "number of connections is unlimited" - by default it's limited in Linux to 1024 per process - someone could make million connections to make server use gigabytes of RAM and crash by using all RAM, but I've never seen anything like that.
If you are interested in configuration of Linux (Ubuntu 20.04) for popular OTS:
tutorials/polish/konfiguracja-serwera-linux at master · gesior/tutorials
Contribute to gesior/tutorials development by creating an account on GitHub.
github.com
There are not just some random commands, everything is described 'why I use this'.
marcelomnzs
Member
- Joined
- Aug 9, 2019
- Messages
- 121
- Reaction score
- 17
I will extract the maximum knowledge from it. thanks master!Only thing that can cause crash is bug in C++ (engine) or Lua (on TFS 1.x where Lua bugs can crash engine).
Of course, some players know engine bugs and abuse them to crash OTSes, to clone items.
In some edge scenario: "number of connections is unlimited" - by default it's limited in Linux to 1024 per process - someone could make million connections to make server use gigabytes of RAM and crash by using all RAM, but I've never seen anything like that.
If you are interested in configuration of Linux (Ubuntu 20.04) for popular OTS:
It's in polish, but I know some people who configured their servers using this with Google Translate.
tutorials/polish/konfiguracja-serwera-linux at master · gesior/tutorials
Contribute to gesior/tutorials development by creating an account on GitHub.
There are not just some random commands, everything is described 'why I use this'.
just to clarify, I'm using ubuntu 22.04, do you think I should use 20.04 to follow these steps?
marcelomnzs
Member
- Joined
- Aug 9, 2019
- Messages
- 121
- Reaction score
- 17
dumped...
the log:
edit:
help pls.. i cant read the logs ;s
the log:
and now? :s
edit:
help pls.. i cant read the logs ;s
Last edited:
pink_panther
Excellent OT User
Do this:
gdb
set logging on
gdb tfs core.85146
bt full
Then attach the log file containing the back trace.
marcelomnzs
Member
- Joined
- Aug 9, 2019
- Messages
- 121
- Reaction score
- 17
I stop here..
searching "lua_pcall" in src and data files i found it:
Similar threads
- Solved
