Re: postgresql 8.4 semaphores ...

[Marc Balmer <marc%msys.ch@localhost>]

Am 07.06.10 22:31, schrieb Manuel Bouyer:
> On Mon, Jun 07, 2010 at 10:24:41PM +0200, Marc Balmer wrote:
>>> I've already seen this. Postgress doesn't destroy IPCs it created when it
>>> exits. ipcs will show you theses, and you can destroy them with ipcrm
>>
>> Is this something we should report/try to change upstream?
> 
> I don't know. Maybe it depends on how postgress is stopped.

I could well imagine that this current behaviour is some sort of defect.
 It does not make much sense let the IPCs hang around after exit, unless
I am mistaken.