I'm not saying you're wrong, but I also think it's true that typical Linux
usage patterns are rather different from those of other *nixen. Linux
started out being able to do a lot with a little, and is still often used
that way - with more functions crammed into boxes with less resources. When
I last worked in a data centre (a few years ago now, for one of the world's
largest companies) they had hundreds of AIX and HP-UX boxes, each well
resourced and each dedicated to exactly one function. I rarely see Linux
being used that way, and I often see it configured with lowish memory and
not nearly enough swap.
In any case, it seems to me we need to have someone check that setting the
vm.overcommit_memory to paranoid will actually stop the postmaster being
killed. I'd love to help but I'm up to my ears in stuff right now. If we
know that we can save the philosophical stuff for another day :-)
cheers
andrew
----- Original Message -----
From: "Tom Lane" <tgl@sss.pgh.pa.us>
To: "Greg Stark" <gsstark@mit.edu>
Cc: <pgsql-hackers@postgresql.org>
Sent: Thursday, June 12, 2003 6:19 PM
Subject: Re: [HACKERS] Pre-allocation of shared memory ...
> Greg Stark <gsstark@mit.edu> writes:
> > I think you'll find this overcommit issue affects many if not most
Unixen.
>
> I'm unconvinced, because I've only ever heard of the problem affecting
> Postgres on Linux.
>
> regards, tom lane