Bug report: je_arena_dalloc_bin_locked () from /usr/lib64/libjemalloc.so.1

[Jason Evans]

On Oct 2, 2014, at 1:40 AM, Roel Van de Paar <roel.vandepaar at percona.com> wrote:
> Hi jemalloc list,
> 
> On Sep 11 I submitted this bug report, but have not seen any updates on it - any progress?

This looks suspiciously like a crash due to the application performing a double free or write after free, but there's not enough information here to know for sure.  You may be able to narrow the problem down by using a debug build, disabling tcache, etc.  Please let us know if you determine that it is in fact a jemalloc bug.

Thanks,
Jason

> On Thu, Sep 11, 2014 at 8:13 AM, Roel Van de Paar <roel.vandepaar at percona.com> wrote:
> Stack (as generated by mysqld):
> 
> Thread 20 (LWP 698):
> +bt
> #0  0x0000000005072771 in pthread_kill () from /lib64/libpthread.so.0
> #1  0x0000000000ad3e5e in my_write_core (sig=11) at /mnt/workspace/percona-server-5.6-binaries-valgrind-yassl/label_exp/centos6-64/percona-server-5.6.20-68.0/mysys/stacktrace.c:422
> #2  0x000000000073156f in handle_fatal_signal (sig=11) at /mnt/workspace/percona-server-5.6-binaries-valgrind-yassl/label_exp/centos6-64/percona-server-5.6.20-68.0/sql/signal_handler.cc:236
> #3  <signal handler called>
> #4  0x0000000004e3ff98 in je_arena_dalloc_bin_locked () from /usr/lib64/libjemalloc.so.1
> #5  0x0000000004e58881 in je_tcache_bin_flush_small () from /usr/lib64/libjemalloc.so.1
> #6  0x0000000004e58fae in je_tcache_destroy () from /usr/lib64/libjemalloc.so.1
> #7  0x0000000004e59183 in je_tcache_thread_cleanup () from /usr/lib64/libjemalloc.so.1
> #8  0x000000000506dbf2 in __nptl_deallocate_tsd () from /lib64/libpthread.so.0
> #9  0x000000000506de01 in start_thread () from /lib64/libpthread.so.0
> #10 0x00000000063dc3dd in clone () from /lib64/libc.so.6
> 
> OS: Centos 7
> 
> jemalloc version: jemalloc.x86_64 0:3.6.0-1.el7 (libjemalloc.so.1 & libpthread.so[.0] attached
> 
> I have also attached a "bundle" of files which we usually attached to mysqld bug reports. 
> It contains a lot of mysqld-related information, like the core dump, stack traces generated etc.). 
> You never know, it may come in handy when researching this bug.
> 
> If you are interested in debug stack traces, please do sent me a compiled debug lib that I can use directly and I'll try to get (gdb) bt full stacks if I can repeat the issue.
> 
> Hope you can fix this. Thanks!

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://jemalloc.net/mailman/jemalloc-discuss/attachments/20141002/390ed890/attachment-0001.html>