On Fri, Nov 7, 2014 at 3:21 PM, <dimon99901@mail.ru> wrote:
> 2014-11-06 13:58:02 GMT LOG: server process (PID 44060) was terminated by
> signal 11: Segmentation fault
> 2014-11-06 13:58:02 GMT DETAIL: Failed process was running: select id_user
> from site.items where id_item = '10795593'
> 2014-11-06 13:58:02 GMT LOG: terminating any other active server processes
> 2014-11-06 13:58:02 GMT WARNING: terminating connection because of crash of
> another server process
> 2014-11-06 13:58:02 GMT DETAIL: The postmaster has commanded this server
> process to roll back the current transaction and exit, because another
> server process exited abnormally and possibly corrupted shared memory.
>
>
> Further, postgres generates 6Kiops for 2min on the HDD (I assume that
> postgres checked data for corruption).
>
> This situation occures 1 times a day for 2 days.
> I planned upgrade to 9.3.5, but i have not seen this bug in 9.3.5 release
> notes.
It is hard to guess what is exactly happening here. The query that is
reported as failing in those logs is a simple one, that it used for
sure millions of times across thousands of application, hence what you
are seeing here may be caused with some other process interactions, or
perhaps something at the hardware level. In any case, a core dump or
more information (did you see unusual errors or things in the past log
entries?) would help to grab an understanding of what is happening.
--
Michael