From owner-freebsd-stable@FreeBSD.ORG Wed Apr 16 22:36:46 2003 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id CEDF337B401 for ; Wed, 16 Apr 2003 22:36:46 -0700 (PDT) Received: from horsey.gshapiro.net (horsey.gshapiro.net [64.105.95.154]) by mx1.FreeBSD.org (Postfix) with ESMTP id 433C543FDD for ; Wed, 16 Apr 2003 22:36:46 -0700 (PDT) (envelope-from gshapiro@gshapiro.net) Received: from horsey.gshapiro.net (localhost [127.0.0.1]) h3H5ajdB087080 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO) for ; Wed, 16 Apr 2003 22:36:45 -0700 (PDT) Received: (from gshapiro@localhost)h3H5ajNb087079 for freebsd-stable@freebsd.org; Wed, 16 Apr 2003 22:36:45 -0700 (PDT) Date: Wed, 16 Apr 2003 22:36:45 -0700 From: Gregory Neil Shapiro To: freebsd-stable@freebsd.org Message-ID: <20030417053645.GD590@horsey.gshapiro.net> References: <200304170312.h3H3C7M2032490@drugs.dv.isc.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200304170312.h3H3C7M2032490@drugs.dv.isc.org> User-Agent: Mutt/1.5.4i Subject: Re: post last sendmail upgrade X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 17 Apr 2003 05:36:47 -0000 On Thu, Apr 17, 2003 at 01:12:07PM +1000, Mark.Andrews@isc.org wrote: > > I've noticed email corruption post the last upgrade to sendmail > committed to stable. Is anyone else experiencing this? I just read through the diff between 8.12.6 (4.7-RELEASE) and 8.12.9 and didn't see anything in sendmail itself that would cause message corruption. I did however spot one oddity in mail.local but older versions would have suffered from the same race condition. Basically, if the close() on the mailbox fails (due to disk full, user over quota, or a write error only caught after the fsync() (which shouldn't happen)), the mailbox is truncated back to the pre-message size without re-locking the mailbox. It's doubtful this is happening to you since the chances of close() failing and the previous fsync() not are pretty low. You would also see temporary delivery failures of the form "450 4.2.0 " in syslog. Are you using mail.local or some other local delivery agent? How are you reading your mail from your mailbox; could that pop/imap/MUA be responsible?