[postgis-users] Signal 11

[Stephen Woodbridge]

Peter,

After reading the thread from nabble, it seems people have overlook a 
possible source for the error, and that would be a memory overwrite that 
just happens to hitting the variable that is set/checked when you get a 
SIGTERM. So without a SIGTERM, if the variable the a SIGTERM sets gets 
overwritten by a bad pointer would simulate a SIGTERM without one 
getting sent.

You should be able to put a watch on the variable in gdb to catch and 
one doing that, then get a back trace on the calling stack to see where 
it came from. So I can't give you the details on how to attach the 
process and watch it because I haven't had to do it in ages, but a 
little research and reading should point the way.

-Steve

On 12/22/2010 2:58 AM, Peter Hopfgartner wrote:
> Could this could be related to
> http://postgresql.1045698.n5.nabble.com/Getting-FATAL-terminating-connection-due-to-administrator-command-td2840596.html.
>
>  This issue is still a major problem for us, but ATM I do not know
> what I can do to get nearer to a solution of it (if Tom Lane can not
> help me, who can?).
>
> Peter Hopfgartner
>
> R3 GIS Srl - GmbH http://www.r3-gis.com
>
>
> --------Paul Ramsey<pramsey at opengeo.org>  wrote-------- Subject: Re:
> [postgis-users] Signal 11 Date: 21.12.2010 17:52
>
>> This sounds similar to a very very very old problem we had when
>> first hooking up GEOS to PostGIS, which was that the C++ exceptions
>> would not be caught, no matter what we did. At some point it seemed
>> like the compilers caught up to what we wanted and things started
>> working. But in the interim period before then we found that we
>> could force it to work by including libstdc++ in the link line when
>> we built PostgreSQL and PostGIS. (LDFLAGS=-lstdc++ ./configure
>> --etc...)
>>
>> The directions for doing that still live on in the README files of
>> some of the older releases (pull 1.1, for example).
>>
>> Might or might not be the problem.
>>
>> Paul
>>
>> On Tue, Dec 21, 2010 at 8:18 AM, Martin
>> Spott<Martin.Spott at mgras.net> wrote:
>>> Hi Nicklas,
>>>
>>> Nicklas Avén wrote:
>>>
>>>> I get this error, but no crash.
>>>
>>> Thanks for checking. Wether it crashes the DB server or not
>>> might probably related to the respective platform or setup - I
>>> guess  ;-)
>>>
>>> Cheers, Martin. -- Unix _IS_ user friendly - it's just selective
>>> about who its friends
>> are !
>>>
>> --------------------------------------------------------------------------
>>>
>>
_______________________________________________
>>> postgis-users mailing list postgis-users at postgis.refractions.net
>>> http://postgis.refractions.net/mailman/listinfo/postgis-users
>>>
>> _______________________________________________ postgis-users
>> mailing list postgis-users at postgis.refractions.net
>> http://postgis.refractions.net/mailman/listinfo/postgis-users
>>
>
> _______________________________________________ postgis-users mailing
> list postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users