Re: BUG #14420: Parallel worker segfault

On Sat, Nov 12, 2016 at 3:36 AM, <rotten(at)windfish(dot)net> wrote:
> The following bug has been logged on the website:
>
> Bug reference: 14420
> Logged by: Rick Otten
> Email address: rotten(at)windfish(dot)net
> PostgreSQL version: 9.6.1
> Operating system: Ubuntu 16.04
> Description:
>
> This just happened on my production 9.6.1 system:
>
> 2016-11-11 21:31:26.215 UTC LOG: worker process: parallel worker for PID
> 1737 (PID 6792) was terminated by signal 11: Segmentation fault
> 2016-11-11 21:31:26.215 UTC LOG: terminating any other active server
> processes
>
> I had a lot of stuff running concurrently, so I'm not sure which query
> caused the failure. There is nothing else obvious or illuminating in the
> logs that I can see. After it recovered, everything went back to normal.
>

It is difficult to find out the problem by above information. Can you
share server log? I am not sure if we can find anything from it, but
at least we can try.

> I figured I should report this even if I don't have much to go on yet.
>

Thanks for the report.

> Is there a core file somewhere that would be of help? (I'm not seeing
> one.)
>

If it is generated it should be in $PGDATA directory. I have found
few articles [1][2] on web related to this which could be useful for
you.

> Should I turn up debug or enable something to catch more information if it
> happens again?
>

I think core file can help here.

[1] - https://wiki.postgresql.org/wiki/Getting_a_stack_trace_of_a_running_PostgreSQL_backend_on_Linux/BSD
[2] - http://bajis-postgres.blogspot.in/2014/03/want-to-ensure-that-postgres-generates.html

--
With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com