Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751094AbWANUkv (ORCPT ); Sat, 14 Jan 2006 15:40:51 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751095AbWANUkv (ORCPT ); Sat, 14 Jan 2006 15:40:51 -0500 Received: from locomotive.csh.rit.edu ([129.21.60.149]:22839 "EHLO locomotive.unixthugs.org") by vger.kernel.org with ESMTP id S1751094AbWANUku (ORCPT ); Sat, 14 Jan 2006 15:40:50 -0500 Message-ID: <43C961D2.90303@suse.com> Date: Sat, 14 Jan 2006 15:40:50 -0500 From: Jeffrey Mahoney User-Agent: Mozilla Thunderbird 1.0.7 (Macintosh/20050923) X-Accept-Language: en-us, en MIME-Version: 1.0 To: Jan Kara Cc: Jan Engelhardt , Linux Kernel Mailing List Subject: Re: reiserfs mount time References: <20060114203410.GB21901@atrey.karlin.mff.cuni.cz> In-Reply-To: <20060114203410.GB21901@atrey.karlin.mff.cuni.cz> X-Enigmail-Version: 0.89.5.0 X-Enigmail-Supports: pgp-inline, pgp-mime Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1627 Lines: 40 Jan Kara wrote: > Hello, > > >>brought to attentino on an irc channel, reiser seems to have the largest >>mount times for big partitions. I see this behavior on at least two >>machines (160G, 250G) and one specially-crafted virtual machine >>(a 1.9TB disk / 1.9TB partition - took somewhere over 120 seconds). >>Here's a dig http://linuxgazette.net/122/misc/piszcz/group001/image002.png >>from http://linuxgazette.net/122/TWDT.html#piszcz >>So, any hint from the reiserfs developers how come reiserfs takes so long? >>Standard mkreiserfs options (none extra passed). > > If I remember correctly, the problem is reiserfs loads bitmaps on mount > and that takes most of the time. Jeff Mahoney has > patches fixing this but I think Hans rejected them because he wants only > bugfixes in reiser3... Yeah, that's the right analysis. ReiserFS caches bitmaps on mount and for large file systems it can take a quite a long time. I've heard reports of up to 15 minutes on multi-TB file systems. As far as these patches getting accepted, Hans was opposed to them initially, but emailed in September saying he decided that they should be accepted after all. These were against 2.6.12.1, and I remerged and split the patches out yesterday evening. I'm going to do a bit of testing and send them out on Monday for inclusion in -mm. -Jeff -- Jeff Mahoney SUSE Labs - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/