*BSD News Article 56373


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!nntp.coast.net!oleane!tank.news.pipex.net!pipex!tube.news.pipex.net!pipex!dish.news.pipex.net!pipex!lade.news.pipex.net!pipex!bt!btnet!demon!news.cs.kiev.ua!monoli!csoft!ktts!aladon!carrier.kiev.ua!info.elvisti.kiev.ua!office.elvisti.kiev.ua!stesin
From: stesin@elvisti.kiev.ua (Andrew V. Stesin)
Newsgroups: comp.unix.internals,comp.unix.admin,comp.unix.bsd.freebsd.misc,relcom.fido.ru.unix
Subject: Re: UN*X cluster on Intel platform?
Date: Fri, 1 Dec 1995 14:11:46 GMT
Organization: Electronni Visti InformAgency (ElVisti)
Lines: 41
Sender: usenet@info.elvisti.kiev.ua (ElVisti USENET)
Message-ID: <DIwvFn.EB9@info.elvisti.kiev.ua>
References: <DIrI2w.FHy@info.elvisti.kiev.ua> <49j9vs$4fu@madeline.INS.CWRU.Edu>
X-NNTP-Posting-Host: bee.cs.kiev.ua
X-Newsreader: TIN [version 1.2 PL2+]
Xref: euryale.cc.adfa.oz.au comp.unix.internals:9418 comp.unix.admin:35792 comp.unix.bsd.freebsd.misc:10029

Hello Wes Brown (wes@prozac.student.cwru.edu); quoting you from relcom.fido.ru.unix <49j9vs$4fu@madeline.INS.CWRU.Edu> : 

: What do you want to do?  {I am sorry VMS was before my time}

: We have LINUX running in what could be considered a cluster configuration.
: One 80386 25MHz is a dedicated machine.  There are other machines that mount
: its disk space.  Those machines are owned by other students that also run DOS
: and other useless excusses for operating systems.

	No, "clustering' means something more than simple NFS mounting.
	I mean that several machines can run (even more than) one task
	in parallel, synchronizing them with some kind of Message Passing
	Interface, with a shared (or divided into several chunks)
	dataset.

	For an example of FreeBSD solution I've recently (shame on me! :) 
	discover see http://www.geli.com and http://www.netlib.org/pvm3

	I wonder if even more sophisticated approach is implementable:
	have several motherbards (each has it's own private disk space,
	too) chained with a single SCSI-2 ribbon cable,
	and a shared disk space on the same cable, too; while messages
	are passed via Fast Ethernet, for example, all access to a shared
	data set is done via a SCSI-2 bus. That might be a very fail-safe
	configuration! :-)

: Wes
: ---
: Wes Brown
: ewb4@po.cwru.edu		wes@prozac.student.cwru.edu
: http://prozac.cwru.edu/wes/About.me.html
: KB8TGR

--

	With best regards -- Andrew Stesin.

	+380 (44) 2760188	+380 (44) 2713457	+380 (44) 2713560

	An undocumented feature is a coding error.