Re: Shared Memory Sizing - Mailing list pgsql-general

From Andrew Sullivan
Subject Re: Shared Memory Sizing
Date
Msg-id 20020627165920.T16498@mail.libertyrms.com
Whole thread Raw
In response to Re: Shared Memory Sizing  ("Glen Parker" <glenebob@nwlink.com>)
Responses Re: Shared Memory Sizing  (Curt Sampson <cjs@cynic.net>)
List pgsql-general
On Thu, Jun 27, 2002 at 01:44:56PM -0700, Glen Parker wrote:

> Obviously a lot of people disagree with this...  I'd like to understand
> why this approach is considered incorrect when postgres rather owns the
> machine?

If you starve other things on the machine for memory, you'll cause
swapping.

Consider all the other things you're doing on the machine -- just
little things, like cron and such.  All that takes memory.
Therefore, it's dangerous not to let the OS manage a good chunk of
memory.

There also appears to be a diminishing returns problem: at a certain
point, you're unlikely to need more shared space, and if you do
something else on the machine that could use the memory, you're
throwing it away.  But I don't see that you're wrong in principle.
Just don't get it wrong -- I _have_ caused a 16 gig machine to swap.
It's not fun.

A

--
----
Andrew Sullivan                               87 Mowat Avenue
Liberty RMS                           Toronto, Ontario Canada
<andrew@libertyrms.info>                              M6K 3E3
                                         +1 416 646 3304 x110




pgsql-general by date:

Previous
From: "Glen Parker"
Date:
Subject: Re: Shared Memory Sizing
Next
From: Patrick Bakker
Date:
Subject: Wrap-around undo/redo log of temporal/time travel/versioned table