Return-Path: <linux-kernel-owner+willy=40w.ods.org@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id <S266367AbTAWXQL>; Thu, 23 Jan 2003 18:16:11 -0500
Received: (majordomo@vger.kernel.org) by vger.kernel.org
	id <S266986AbTAWXQL>; Thu, 23 Jan 2003 18:16:11 -0500
Received: from smtp-out-4.wanadoo.fr ([193.252.19.23]:11179 "EHLO
	mel-rto4.wanadoo.fr") by vger.kernel.org with ESMTP
	id <S266367AbTAWXQJ>; Thu, 23 Jan 2003 18:16:09 -0500
Subject: Re: Expand VM
From: Thomas Cataldo <tomc@compaqnet.fr>
To: User & <breno_silva@beta.bandnet.com.br>
Cc: Valdis.Kletnieks@vt.edu, linux-kernel@vger.kernel.org
In-Reply-To: <20030123194014.M374@beta.bandnet.com.br>
References: <20030123155627.M95099@beta.bandnet.com.br>
	 <200301231655.h0NGtc75010414@turing-police.cc.vt.edu>
	 <20030123194014.M374@beta.bandnet.com.br>
Content-Type: text/plain; charset=ISO-8859-1
Organization: 
Message-Id: <1043364034.25529.18.camel@localhost>
Mime-Version: 1.0
X-Mailer: Ximian Evolution 1.2.1 
Date: 24 Jan 2003 00:20:34 +0100
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
X-Mailing-List: linux-kernel@vger.kernel.org

On Thu, 2003-01-23 at 20:40, User & wrote:
> Hi Valdis
> 
> Create a new VMA on Linux B for Linux A is easy , but i have a problem , the 
> address of VMA is returned on Linux B , so the VMA created on Linux B can not 
> be used for process of linux A.
> The problem is "how can i return address of VMA created on LINUX B to Linux 
> A , and use this space ?".
> 


What you're looking for is openmosix. It does process migration and so
on..

> Thanks
> Breno
> 
> On Thu, 23 Jan 2003 11:55:38 -0500, Valdis.Kletnieks wrote
> > On Thu, 23 Jan 2003 12:56:27 -0300, User & 
> > <breno_silva@beta.bandnet.com.br>  said:
> > 
> > > I have one idea , and this is about expand virtual memory on linux boxes 
> > > connected in LAN.
> > > Example: Linux A is processing come information , and need more memory , 
> so 
> > > with this source , Linux A could access virtual memory on Linux B in LAN.
> > 
> > We've seen *this* done before (remember diskless Sun3-50's?) - the /dev/swap
> > file would be a large file on an NFS mount from a server.  At the 
> > time, this actually made performance sense, because the old 
> > 'Shoebox' drives the -50 came with were incredibly slow, and you 
> > could actually do an NFS operation to a larger server (a -280 with 
> > Fujitsu SuperEagle disks, for instance) faster than talking to the 
> > local disk.
> > 
> > These days, it's probably easier and cheaper to just buy more RAM 
> > and/or disk for Linux A.
> > 
> > > But i don?t know how translate the virtual address between Linux A and 
> B , to 
> > > have success in acess VM, or how to send all the process for Linux B to 
> be 
> > > processed.
> > 
> > Sending the whole process to Linux B to be processed is called "process
> > migration", and is a difficult problem.  Moving the memory image of the
> > process is usually pretty easy.  What is difficult is moving things like
> > references to open files, file locks, and so on (what if the process 
> > is actively writing to block 739 of /usr/foo/some.file, and the 
> > LinuxB machine doesn't have a /usr/foo, or the permissions on 
> > some.file don't match, or another process has it locked, or... ) 
> > There be nasty dragons in this.
> > 
> > You're probably better off buying more RAM and disk for your A machine.
> > -- 
> > 				Valdis Kletnieks
> > 				Computer Systems Senior Engineer
> > 				Virginia Tech
> 
> 
> 
> ----------------------
> WebMail Bandnet.com.br
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
-- 
Thomas Cataldo <tomc@compaqnet.fr>

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/