From owner-freebsd-hackers  Wed Feb 19 05:58:02 1997
Return-Path: <owner-hackers>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id FAA23820
          for hackers-outgoing; Wed, 19 Feb 1997 05:58:02 -0800 (PST)
Received: from wgold.demon.co.uk (wgold.demon.co.uk [158.152.96.124])
          by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id FAA23808
          for <hackers@FreeBSD.ORG>; Wed, 19 Feb 1997 05:57:57 -0800 (PST)
Received: from wgold.demon.co.uk by wgold.demon.co.uk (NTMail 3.02.10) with ESMTP id sa001188 for <hackers@FreeBSD.ORG>; Wed, 19 Feb 1997 11:22:39 +0000
Message-ID: <330AE27F.7159@wgold.demon.co.uk>
Date: Wed, 19 Feb 1997 11:22:39 +0000
From: James Mansion <james@wgold.demon.co.uk>
Organization: Westongold Ltd
X-Mailer: Mozilla 3.0Gold (WinNT; I)
MIME-Version: 1.0
To: hackers@FreeBSD.ORG
Subject: [Fwd: Failed mail: unknown user]
Content-Type: multipart/mixed; boundary="------------69D23902C26"
X-Info: Westongold Ltd: +44 1992 620025 www.westongold.com
Sender: owner-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

This is a multi-part message in MIME format.

--------------69D23902C26
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

(Damn, finger trouble)

--------------69D23902C26
Content-Type: message/rfc822
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

Received: from wgold.demon.co.uk by wgold.demon.co.uk (NTMail 3.02.10) with ESMTP id james for <james>; Wed, 19 Feb 1997 11:21:09 +0000
From: "root@wgold.demon.co.uk" <root@wgold.demon.co.uk>
To: "james@wgold.demon.co.uk" <james@wgold.demon.co.uk>
Subject: Failed mail: unknown user
Date: Wed, 19 Feb 1997 11:21:09 +0000
Return-Path: <>
Message-Id: <11210972800008@wgold.demon.co.uk>
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="==11210975800009=="

This is a MIME-encapsulated message

--==11210975800009==
Content-Type: text/plain; charset=us-ascii

This mail message has been delivered to the default mailbox.
The requested destination was:
   FreeBSD.ORG@wgold.demon.co.uk
 
The text of the message follows:
 

--==11210975800009==
Content-Type: message/rfc822

Message-ID: <330AE225.77E7@wgold.demon.co.uk>
Date: Wed, 19 Feb 1997 11:21:09 +0000
From: James Mansion <james@wgold.demon.co.uk>
Organization: Westongold Ltd
X-Mailer: Mozilla 3.0Gold (WinNT; I)
MIME-Version: 1.0
To: hackers:FreeBSD.ORG
Subject: Disk flush
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Info: Westongold Ltd: +44 1992 620025 www.westongold.com

I'm trying to build a transactional data manager.

Assume that I'm using a raw device and I set it up so that
I flush through the caches.  (Yeah, this is imprecise, but
the precise details are not important, and besides are
different for some platforms)

My question (admitedly not totally BSD related):

 - if the OS tells me it has a successful write through the
   disk device, can I rely on:

   a) the data is on the platter
   b) the data is in the drives buffer and may be on the
      platter

 - if the system crashes during the write (literally, while
   the drive is writing to the platter) can I:

   a) rely on a sector being there, or not (I doubt this on most
      hardware)
   b) rely on the data that was not yet written (perhaps a part-
      sector) being uncorrupted?

My plan is to flush everyting out, then perform a second write
in which just one bit is changed - and I'm hoping that this single
change will be atomic.

Anyone got any insights?

James

--==11210975800009==--

--------------69D23902C26--