From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <6f2d88478639cf5b64080093392c4a1e@quanstro.net>
From: erik quanstrom <quanstro@quanstro.net>
Date: Sun, 17 Sep 2006 16:18:11 -0500
To: 9fans@cse.psu.edu
Subject: Re: [9fans] install "out of physical memory"
In-Reply-To: <775b8d190609171410q4dacb30bm2ed211fcf87c821b@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Topicbox-Message-UUID: b93cbff2-ead1-11e9-9d60-3106f5b1d025

let's face it.  there are some questions that are best left unasked.  buy
lots of memory, keep a swap partition and keep your fingers crossed
that you never need an answer to that question --- what do i do when
i run out of memory!

either that, or do like knuth.  we keep a fixed tablesize for everything.
we size these variables to the memory size of the machine, then we
will never run out of memory.  if you try to start one too many processes,
it will helpfully give an error message before you start. ;-)

- erik

On Sun Sep 17 16:11:27 CDT 2006, bruce.ellis@gmail.com wrote:
> i've never had problems with swap.  my setup at the labs was diskless
> "terminal" with lotsa ram and swap on /n/other.  it never failed, tho swap
> was really only there to cope with extraordinary circumstances.
> 
> i have a similar setup in sydney tho "other" is local.
> 
> killing a random process is not a solution.  it's an egregious hack.
> 
> brucee
>