Random DB core, thr_kill.S

[Nek0. 0]

Hello guys, I have this random crash core that happens every 1-2-3 days randomly on my server:

 

Quote

(gdb) file db
Reading symbols from db...
(gdb) core db.core
[New LWP 104209]
[New LWP 101402]
[New LWP 104210]
[New LWP 104211]
[New LWP 104212]
[New LWP 104213]
[New LWP 104214]
[New LWP 104215]
[New LWP 104216]
Core was generated by `srv1-db'.
Program terminated with signal SIGABRT, Aborted.
Sent by thr_kill() from pid 16013 and user 0.
#0  thr_kill () at thr_kill.S:4
4       thr_kill.S: No such file or directory.
[Current thread is 1 (LWP 104209)]
(gdb) bt full
#0  thr_kill () at thr_kill.S:4
No locals.
#1  0x0834237a in __raise (s=6) at /usr/src/lib/libc/gen/raise.c:52
        id = <optimized out>
#2  0x0837a57d in abort () at /usr/src/lib/libc/stdlib/abort.c:67
        act = {__sigaction_u = {__sa_handler = 0x0, __sa_sigaction = 0x0},
          sa_flags = 0, sa_mask = {__bits = {4294967263, 4294967295, 4294967295,
              4294967295}}}
#3  0x0804c44c in uw_init_context_1 (context=context@entry=0xfbbfde68,
    outer_cfa=outer_cfa@entry=0xfbbfdf88, outer_ra=0x833134a <thread_unwind+58>)
    at /wrkdirs/usr/ports/lang/gcc11/work/gcc-11.3.0/libgcc/unwind-dw2.c:1593
        ra = <optimized out>
        fs = {regs = {reg = {{loc = {reg = 0, offset = 0, exp = 0x0},
                how = REG_UNSAVED} <repeats 18 times>}, prev = 0x0,
            cfa_offset = 0, cfa_reg = 0, cfa_exp = 0x0, cfa_how = CFA_UNSET},
          pc = 0x0, personality = 0x0, data_align = 0, code_align = 0,
          retaddr_column = 0, fde_encoding = 0 '\000', lsda_encoding = 0 '\000',
          saw_z = 0 '\000', signal_frame = 0 '\000', eh_ptr = 0x0}
        sp_slot = {ptr = 0, word = 0}
        code = <optimized out>
#4  0x0832d4e9 in _Unwind_ForcedUnwind (exc=0x28738590,
    stop=0x8331510 <thread_unwind_stop>, stop_argument=0x0)
    at /wrkdirs/usr/ports/lang/gcc11/work/gcc-11.3.0/libgcc/unwind.inc:211
        this_context = {reg = {0x0 <repeats 18 times>}, cfa = 0x0,
          ra = 0x832d4e9 <_Unwind_ForcedUnwind+41>, lsda = 0x0, bases = {
            tbase = 0x0, dbase = 0x0, func = 0x0}, flags = 1073741824,
          version = 0, args_size = 0, by_value = '\000' <repeats 17 times>}
--Type <RET> for more, q to quit, c to continue without paging--c
        cur_context = {reg = {0x0, 0xfbbfdf24,
            0x8335d06 <mutex_unlock_common+198>, 0x0, 0xfbbfdf38,
            0x80f7e7a <CAsyncSQL::CopyQuery()+346>, 0x0, 0x0, 0x1, 0x19711,
            0x287385f0, 0x0, 0x2873d804, 0x28763c00, 0x2ae5e040, 0xfbbfdf38,
            0x8335b7d <_thr_mutex_unlock+29>, 0x0}, cfa = 0x28737f00,
          ra = 0x28763c00, lsda = 0xfbbfdf98, bases = {
            tbase = 0x80f86c6 <CAsyncSQL::ChildLoop()+2006>, dbase = 0x2872c0f8,
            func = 0x0}, flags = 4223655776, version = 142474760,
          args_size = 678659032,
          by_value = "x߿\373\000\000\000\000`\340\345*@\340\345* \241"}
        code = <optimized out>
        frames = 15
#5  0x0833134a in thread_unwind () at /usr/src/lib/libthr/thread/thr_exit.c:175
        curthread = 0x28738400
#6  0x083312b2 in _pthread_exit_mask (status=0x0, mask=0x0)
    at /usr/src/lib/libthr/thread/thr_exit.c:257
        curthread = 0x28738400
        cleanup = <optimized out>
#7  0x0833121d in _Tthr_exit (status=0x0)
    at /usr/src/lib/libthr/thread/thr_exit.c:209
No locals.
#8  0x08330ea0 in thread_start (curthread=0x28738400)
    at /usr/src/lib/libthr/thread/thr_create.c:292
        set = {__bits = {0, 0, 0, 0}}
#9  0x00000000 in ?? ()
No symbol table info available.
(gdb)
 

I tried to include everything inside gdb, please let me know if you guys have any idea why this is happening. Freebsd 13.1 i386 gen1 v1.

To mention that the source is compiled and runned on the same server, no differences in Freebsd versions.

[Intel 734]

Question: do you, perhaps, use gcc? And does this core happens even when just shutting down the game?

Because I used to have this whenever I shut down the cores:

 

Never when online though. Recently I changed to clang17 on FreeBSD 13.2 and.. this does not happen.

I would also add: same code, with freebsd 11.1 I never had these problems. Since 12.2 I had started having this.

Edited November 4, 2023 by Metin2 Dev International
Core X - External 2 Internal

[Nek0. 0]

1 hour ago, Intel said:

Question: do you, perhaps, use gcc? And does this core happens even when just shutting down the game?

Because I used to have this whenever I shut down the cores:

 

Never when online though. Recently I changed to clang17 on FreeBSD 13.2 and.. this does not happen.

I would also add: same code, with freebsd 11.1 I never had these problems. Since 12.2 I had started having this.

Ohh, you're right! It only happens when I shut down the server, so it shouldn't be a problem when it's live.Have any tutorial to adapt marty's source to clang17?

[Intel 734]

4 minutes ago, Nek0. said:

Ohh, you're right! It only happens when I shut down the server, so it shouldn't be a problem when it's live.Have any tutorial to adapt marty's source to clang17?

If you are his customer his latest files are compatible with both clang and gcc (and you can switch with a single script) (In general, switch the compiler in the makefiles and follow the errors. You'd need to recompile some libs, like cryptopp. DevIL maybe, format for sure if you use it)

 

Btw I never had issues with this core crash, but I wouldn't bet on it to leave it like this ahah