[zeromq-dev] my day for oddities

Dhammika Pathirana dhammika at gmail.com
Wed Oct 27 07:19:57 CEST 2010


Hi Andrew,


On Tue, Oct 26, 2010 at 7:06 AM, Andrew Hume <andrew at research.att.com> wrote:
> i am getting sporadic core dumps with my zmq programs.
> no pattern detectable yet, except that whenever it happens,
> the immediate cause is a zero zmq_msg_t. e.g.
> (gdb) bt
> #0  zmq_msg_close(._0 *) (msg_=0x7fff1e481d10) at zmq.cpp:171
> #1  0x00002aefdbe7eff8 in zmq::pipe_t::~pipe_t (this=0x32beb30,
>     __in_chrg=<value optimized out>) at pipe.cpp:333
> #2  0x00002aefdbe7c21b in zmq::object_t::process_command (this=0x32beb90,
>     cmd_=...) at object.cpp:102
> #3  0x00002aefdbe7449c in zmq::app_thread_t::process_commands (
>     this=0x31af0f0, block_=<value optimized out>,
>     throttle_=<value optimized out>) at app_thread.cpp:128
> #4  0x00002aefdbe8d384 in zmq_poll(._1 *, int, long) (items_=0x7fff1e481fc0,
>     nitems_=2, timeout_=-1) at zmq.cpp:448
> #5  0x0000000000401a00 in main (argc=0, argv=0x7fff1e48a790) at kahuna.c:152
> (gdb) print *msg_
> $2 = {content = 0x0, flags = 0 '\000', vsm_size = 0 '\000',
>   vsm_data = '\000' <repeats 29 times>}
> (gdb) quit
> as you can see, this is happening deep inside zmq_poll, but i have also seen
> this zero msg being returned by zmq_recv.
> i am running 2.0.9 on red hat enterprise 5.4.



Looks like ypipe issue that was fixed in c1deb22, but it's in 2.0.10.
Can you give it a try? It's a trivial fix, if this is in production we
can merge it into 2.0.9.


Dhammika



More information about the zeromq-dev mailing list