[rabbitmq-discuss] RabbitMQ Causing Segmentation Fault on FreeBSD

[Brandon Munroe]

Here's the output of the command around (within 10s) when it segfaults.
I've truncated a bunch of similar lines, since there are more than 25k
lines.

USER     CMD          PID   FD MOUNT      INUM MODE         SZ|DV R/W
brandon  beam.smp   65462 text /        17977559 -r-xr-xr-x  2397040  r
brandon  beam.smp   65462 ctty /dev         89 crw--w----   pts/0 rw
brandon  beam.smp   65462   wd /        15810621 drwxr-xr-x    1024  r
brandon  beam.smp   65462 root /             2 drwxr-xr-x    1024  r
brandon  beam.smp   65462    0 /dev         89 crw--w----   pts/0 rw
brandon  beam.smp   65462    1 /dev         89 crw--w----   pts/0 rw
brandon  beam.smp   65462    2 /dev         89 crw--w----   pts/0 rw
brandon  beam.smp   65462    3
brandon  beam.smp   65462    4* pipe fffff8000a978000 <->
fffff8000a978160      0 rw
brandon  beam.smp   65462    5* pipe fffff8000a978160 <->
fffff8000a978000      0 rw
brandon  beam.smp   65462    6* pipe fffff8008b5b22e8 <->
fffff8008b5b2448      0 rw
brandon  beam.smp   65462    7* pipe fffff8008b5b2448 <->
fffff8008b5b22e8      0 rw
brandon  beam.smp   65462    8* internet stream tcp fffff8027d821000
brandon  beam.smp   65462    9* internet stream tcp fffff80566596800
brandon  beam.smp   65462   10 -        15810857 -rw-r--r--  3093764  w
brandon  beam.smp   65462   11 -        15810858 -rw-r--r--       0  w
brandon  beam.smp   65462   12 -        15891126 -rw-r--r--      96 rw
brandon  beam.smp   65462   13 -        15891137 -rw-r--r--       0 rw
brandon  beam.smp   65462   14 -        15891139 -rw-r--r--       0 rw
brandon  beam.smp   65462   15* internet6 stream tcp fffff80458487800
brandon  beam.smp   65462   16* pipe fffff803c3f4b000 <->
fffff803c3f4b160      0 rw
brandon  beam.smp   65462   17* internet stream tcp fffff80458449400
brandon  beam.smp   65462   18* internet stream tcp fffff802be9b4c00
brandon  beam.smp   65462   19* pipe fffff8000a9bca18 <->
fffff8000a9bc8b8      0 rw
brandon  beam.smp   65462   20* internet stream tcp fffff802bee95400
brandon  beam.smp   65462   21* internet stream tcp fffff802be4ad400
brandon  beam.smp   65462   22* pipe fffff801ea5f52e8 <->
fffff801ea5f5448      0 rw
brandon  beam.smp   65462   23* internet stream tcp fffff804edad0800
brandon  beam.smp   65462   24* internet stream tcp fffff80249ff8400
brandon  beam.smp   65462   25* internet stream tcp fffff80566026400
...
brandon  beam.smp   65462 25018* internet stream tcp fffff8045848d000
brandon  beam.smp   65462 25019* internet stream tcp fffff805662fb400
brandon  beam.smp   65462 25020* internet stream tcp fffff804a4633400
brandon  beam.smp   65462 25021* internet stream tcp fffff802495a8000
brandon  beam.smp   65462 25024* pipe fffff8049522a160 <->
fffff8049522a000      0 rw

Let me know if you need anything else, or if you spot anything that I can
do to mitigate the issue.

-Brandon


On Fri, Apr 18, 2014 at 8:16 AM, Matthias Radestock
<matthias at rabbitmq.com>wrote:

> Brandon,
>
>
> On 15/04/14 17:25, Brandon Munroe wrote:
>
>> The message in the logs reads:
>>
>> ** Generic server rabbit_mgmt_external_stats terminating
>> ** Last message in was emit_update
>> ** When Server state == {state,707130}
>> ** Reason for termination ==
>> ** {function_clause,[{lists,nth,[1,[]],[{file,"lists.erl"},{line,168}]},
>>
>> {rabbit_mgmt_external_stats,'-get_used_fd/1-fun-1-',2,[]},
>>                       {lists,'-filter/2-lc$^0/1-0-',2,
>>                              [{file,"lists.erl"},{line,1271}]},
>>                       {lists,'-filter/2-lc$^0/1-0-',2,
>>                              [{file,"lists.erl"},{line,1271}]},
>>                       {rabbit_mgmt_external_stats,get_used_fd,1,[]},
>>
>> {rabbit_mgmt_external_stats,'-infos/2-lc$^0/1-0-',2,[]},
>>
>> {rabbit_mgmt_external_stats,'-infos/2-lc$^0/1-0-',2,[]},
>>                       {rabbit_mgmt_external_stats,emit_update,1,[]}]}
>>
>
> On BSD RabbitMQ invokes 'fstat -p <pid>' to determine the number of used
> file descriptors. In the above it is tripping over the parsing of the
> output. If you can, please post the output of that command just before
> rabbit dies. Meanwhile I've filed a bug to make our parsing code more
> robust.
>
>
> Matthias.
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20140418/e8d8ab9a/attachment.html>