Fatal Error leads to segmentation fault.

[trickzjen23]

Need help with the DOS. I'm getting this fatal error message when I could no longer log in to the SSH, but when i have an active SSH, the response when typing is very slow.

 

[Fatal Error]: Server has crashed without a connection to the char-server, 61 characters can't be saved!

Segmentation fault.

 

Before the error above happen, I would not be able to login to the ssh. Once I had an active ssh, the char_server_sql is consuming unusual usage of the memory 50% and so as the SQL

 

I already installed CSF limiting access to UDP and TCP ports, and contacting my HOST. My host said it couldn't be a DDOS attack.

 

 

[anacondaq]

https://github.com/rathena/rathena/wiki/Crash_Logging

https://github.com/rathena/rathena/wiki/GDB

https://rathena.org/board/topic/53447-how-to-enable-crashcore-dumps-in-your-server-linux/

https://rathena.org/board/topic/63493-how-to-install-gdb-in-centos/

 

tl;tr you should install gdb debugger for your system and compile your server with enabled debug mode. Then run server again and wait for next crash, when crash will hapaned you will get core file from your server (char, login or map (which were crashed)). Then, with GDB you should try to trace this dump and you will get information about crash (steps were have been done before crash) in your gdb, then these data with crash reason you should to send to issue tracker. But before sending, be sure that you are using LATEST rathena hash (revision), because a lot of crashes already fixed in commits regularly. Even if you don't using rAthena, or using modified rAthena emulator, you can see in GDB output the line in your code which provoke a crash, and starting from this step you can analyze by yourself which thing provoke a crash and how to fix it.

[trickzjen23]

Here is the BT Full from the coredump.

Quote

#0  0x0000000000430803 in chrif_authreq (sd=0x234c8db0) at chrif.c:564
        node = <value optimized out>
#1  0x000000000044c414 in clif_parse (fd=44) at clif.c:15053
        cmd = <value optimized out>
        packet_ver = <value optimized out>
        packet_len = 19
        err = 0
        sd = 0x0
        pnum = <value optimized out>
#2  0x0000000000533eec in do_sockets (next=<value optimized out>)
    at socket.c:929
        rfd = {__fds_bits = {8745810087112071040, 226, 0 <repeats 14 times>}}
        timeout = {tv_sec = 0, tv_usec = 50000}
        ret = <value optimized out>
        i = <value optimized out>
#3  0x00000000005316e5 in main (argc=1, argv=0x7fffac036f38) at core.c:265
        next = <value optimized out>

 
Can anyone help to prevent the map server to crash?
 

 

bump

Edited November 15, 2016 by trickzjen23