From owner-freebsd-questions  Fri Sep 12 19:15:15 1997
Return-Path: <owner-freebsd-questions>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id TAA26585
          for questions-outgoing; Fri, 12 Sep 1997 19:15:15 -0700 (PDT)
Received: from dan.emsphone.com (dan@dan.emsphone.com [199.67.51.101])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id TAA26580
          for <freebsd-questions@FreeBSD.ORG>; Fri, 12 Sep 1997 19:15:13 -0700 (PDT)
Received: (from dan@localhost)
	by dan.emsphone.com (8.8.6/8.8.6) id VAA25756;
	Fri, 12 Sep 1997 21:15:04 -0500 (CDT)
Message-ID: <19970912211504.29147@emsphone.com>
Date: Fri, 12 Sep 1997 21:15:04 -0500
From: Dan Nelson <dnelson@emsphone.com>
To: spork <spork@super-g.com>
Cc: freebsd-questions@FreeBSD.ORG
Subject: Re: Clustering/fail-over capability?
References: <341965AD.7688@PartsNow.com> <Pine.BSF.3.96.970912211308.12209A-100000@super-g.inch.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.82e
In-Reply-To: <Pine.BSF.3.96.970912211308.12209A-100000@super-g.inch.com>; from "spork" on Fri Sep 12 21:15:14 GMT 1997
X-OS: FreeBSD 2.2-970701-RELENG
Sender: owner-freebsd-questions@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

In the last episode (Sep 12), spork said:
> To stray slightly from this thread, has anyone seen this type of
> setup in the PeeCee world:
> 
> SCSI-ADAPTER----DRIVE----DRIVE----SCSI-ADAPTER
> 
> I've seen mention that two controllers can exist on the same chain,
> is that true?  What type of things could one do with this?
> 
> Off to read the SCSI-FAQ,
> 
> Charles

It's certainly possible.  I did a few years ago, with:

  T--HD(0)---Amiga(7)----------PC(6)---HD(1)--T

  |<----Machine 1---->|   |<----Machine 2---->|
      boots from (0)          boots from (1)

(Numbers in parens are the SCSI IDs).  I recall having to be careful
booting up the Amiga, as it didn't like probing the other host adapter.

And I'm doing it right now at work:

T--HD(0)--HD(1)--PC(7)-------RAID(4)(5)-------PC(6)--HD(3)--T

|<-----Machine 1------>|  |<-Raid array->|  |<---Machine 2--->|
     boots from (0)                            boots from (3)

Each machine has a 60-gig partition on the RAID array, and NFS mounts
the other over full-duplex 100mbit Ethernet.  I'm not doing any
failover (the machines are reliable, and aren't in use after working
hours), but I conveivably could determine that the other server is
down, fsck it's disk, and mount it on the other machine.


	-Dan Nelson
	dnelson@emsphone.com