From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261999AbUCVO14 (ORCPT ); Mon, 22 Mar 2004 09:27:56 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262005AbUCVO14 (ORCPT ); Mon, 22 Mar 2004 09:27:56 -0500 Received: from mail2-116.ewetel.de ([212.6.122.116]:54488 "EHLO mail2.ewetel.de") by vger.kernel.org with ESMTP id S261999AbUCVO1z (ORCPT ); Mon, 22 Mar 2004 09:27:55 -0500 Date: Mon, 22 Mar 2004 15:27:31 +0100 (CET) From: Pascal Schmidt To: Stefan Smietanowski cc: Chris Friesen , Chris Wedgwood , Frank Cusack , linux-kernel@vger.kernel.org Subject: Re: Does Linux sync(2) wait? In-Reply-To: <405ED755.2070301@stesmi.com> Message-ID: References: <1C8xa-5lk-5@gated-at.bofh.it> <20040322005953.GA12237@dingdong.cryptoapps.com> <405E611D.10008@nortelnetworks.com> <405ED755.2070301@stesmi.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-CheckCompat: OK Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 22 Mar 2004, Stefan Smietanowski wrote: > >> 20 minutes?! > > He did say it was a magneto-optical drive. > That's 400KiB/s you know - pretty slow. Yes. It's almost exclusively used once per week for backup purposes. Speed doesn't matter for that (about 400 MB of often changing data on /home), reliability counts. Data that doesn't really change (old kernel releases and such) is also on MO, but that's "write once, than forget about it", so once again speed doesn't matter that much. -- Ciao, Pascal