Return-Path: Received: from bombadil.infradead.org ([18.85.46.34]:56160 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932346Ab0BDPVt (ORCPT ); Thu, 4 Feb 2010 10:21:49 -0500 Date: Thu, 4 Feb 2010 10:21:47 -0500 From: Christoph Hellwig To: Trond Myklebust Cc: Ben Myers , linux-nfs@vger.kernel.org Subject: Re: [RFC PATCH 0/4] wsync export option Message-ID: <20100204152147.GB22014@infradead.org> References: <20100203233755.17677.96582.stgit@case> <1265241511.2632.12.camel@localhost> Content-Type: text/plain; charset=us-ascii In-Reply-To: <1265241511.2632.12.camel@localhost> Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 On Wed, Feb 03, 2010 at 06:58:31PM -0500, Trond Myklebust wrote: > Why should the administrator have to both change /etc/fstab > and /etc/exports? That will be an immediate source of trouble if someone > changes one without changing the other. I agree, this should be left to the filesystems. > Why not rather add an optional operation to the export_ops to let the > filesystem specify exactly what kind of synchronisation policy is > optimal for it? Maybe. But I suspect fsync is actually the right thing to do for all the filesystems. The only reason why we'd still might want to make it an export operation is that I'd really prefer not to grow more instances of fsync without the file pointer. Then again they're already all in nfs and it doesn't matter too much to audit a few more places later.