Re: checking on buildfarm member thrush

Gaetano Mendola <mendola(at)bigfoot(dot)com> writes:
> Tom Lane wrote:
>> I'm trying to determine why thrush has been failing on PG CVS HEAD for
>> the past few days. Could you try running the attached program on that
>> machine, and see what it prints? I suspect it will dump core :-(

> $ gcc -D_GNU_SOURCE tom.c -o tom
> $ ./tom
> failed: Success

Hmph. Well, that destroys the posix_fadvise-dumps-core theory, and
leaves us still wondering why thrush is failing "make check".
There's been a core dump recently introduced *somewhere*:

NOTICE: type reference hobbies_r.person%TYPE converted to text
LOG: server process (PID 2818) was terminated by signal 11
LOG: terminating any other active server processes

Could you get a gdb stack trace from that crash? If the buildfarm
run is under a suitable ulimit, it should be leaving a core file
in the test PGDATA directory.

regards, tom lane