From owner-freebsd-hubs@FreeBSD.ORG Thu Jul 8 10:47:22 2010 Return-Path: Delivered-To: hubs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 6F173106566B for ; Thu, 8 Jul 2010 10:47:22 +0000 (UTC) (envelope-from olli@lurza.secnetix.de) Received: from lurza.secnetix.de (lurza.secnetix.de [IPv6:2a01:170:102f::2]) by mx1.freebsd.org (Postfix) with ESMTP id E4EDA8FC0A for ; Thu, 8 Jul 2010 10:47:21 +0000 (UTC) Received: from lurza.secnetix.de (localhost [127.0.0.1]) by lurza.secnetix.de (8.14.3/8.14.3) with ESMTP id o68AkkXt083903; Thu, 8 Jul 2010 12:47:02 +0200 (CEST) (envelope-from oliver.fromme@secnetix.de) Received: (from olli@localhost) by lurza.secnetix.de (8.14.3/8.14.3/Submit) id o68AkiUw083900; Thu, 8 Jul 2010 12:46:44 +0200 (CEST) (envelope-from olli) From: Oliver Fromme Message-Id: <201007081046.o68AkiUw083900@lurza.secnetix.de> To: cejkar@fit.vutbr.cz (Cejka Rudolf) Date: Thu, 8 Jul 2010 12:46:44 +0200 (CEST) In-Reply-To: <20100708083522.GA89496@fit.vutbr.cz> X-Mailer: ELM [version 2.5 PL8] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.3.4 (lurza.secnetix.de [127.0.0.1]); Thu, 08 Jul 2010 12:47:02 +0200 (CEST) Cc: Garrett Wollman , hubs@freebsd.org Subject: Re: FreeBSD 8.1-RC2 Available... X-BeenThere: freebsd-hubs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: "FreeBSD Distributions Hubs: mail sup ftp" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 08 Jul 2010 10:47:22 -0000 Cejka Rudolf wrote: > I'm not ZFS expert either, but I have similar experiences and I rather > stayed with UFS2. You can improve ZFS performance; there are some tuning guides floating around. But I think you will still not reach the performance of a well-tuned UFS. So it's propably better indeed to stay with UFS if performance is the most important aspect of the file system. > I think that long fsck is not such a big problem for > our mirrors and I can see on my 8 x 500 GB HW RAID-5 sequential reads > up to 300 MB/s using UFS2, instead of just 30 MB/s using ZFS (tested > both over RAID-5 and JBOD). If fsck time is an issue, I recommend using gjournal and put the journal on a separate fast disk, e.g. an SSD. It won't affect read performance at all, and reduce write performance only slightly, depending on access patterns. > The other problems are limited capabilities using IPv6 and I really > do not believe in identity of source and destination trees mirrored > using cvsup... To be honest, I trust cvsup more than I trust rsync. I've had some horrible experiences with rsync and lost some backups in the process; especially its error-handling seems to be very bad. For example, when the target file system is full, all kinds of weird things happen, such as files being moved to the wrong directory, plain files being replaced by device nodes and similar oddities. Well, maybe rsync was improved in the meantime; I didn't dare to touch it in a while. :-) Just my 2 cents. Best regards Oliver -- Oliver Fromme, secnetix GmbH & Co. KG, Marktplatz 29, 85567 Grafing b. M. Handelsregister: Registergericht Muenchen, HRA 74606, Geschäftsfuehrung: secnetix Verwaltungsgesellsch. mbH, Handelsregister: Registergericht Mün- chen, HRB 125758, Geschäftsführer: Maik Bachmann, Olaf Erb, Ralf Gebhart FreeBSD-Dienstleistungen, -Produkte und mehr: http://www.secnetix.de/bsd "The most important decision in [programming] language design concerns what is to be left out." -- Niklaus Wirth