From mboxrd@z Thu Jan 1 00:00:00 1970 From: klink@clouddancer.com (Colonel) Subject: (unknown) Date: Tue, 4 Jun 2002 08:47:12 -0700 (PDT) Sender: linux-raid-owner@vger.kernel.org Message-ID: <20020604154712.AD48D8347@phoenix.clouddancer.com> Return-path: To: linux-raid@vger.kernel.org, roy@karlsbakk.net List-Id: linux-raid.ids From: Colonel To: roy@karlsbakk.net CC: linux-raid@vger.kernel.org In-reply-to: <200206041259.g54CxuP07700@mail.pronto.tv> (message from Roy Sigurd Karlsbakk on Tue, 4 Jun 2002 14:59:55 +0200) Subject: Re: SV: RAID-6 support in kernel? Reply-to: klink@clouddancer.com References: <2D0AFEFEE711D611923E009027D39F2B02F17E@nasexs1.meridian-data.com> <200206041259.g54CxuP07700@mail.pronto.tv> From: Roy Sigurd Karlsbakk Organization: Pronto TV AS Date: Tue, 4 Jun 2002 14:59:55 +0200 Cc: Christian Vik , linux-kernel@vger.kernel.org, linux-raid@vger.kernel.org Sender: linux-raid-owner@vger.kernel.org X-Mailing-List: linux-raid@vger.kernel.org News-Group: list.kernel > Of course, for a 4 drive setup there's no reason to use RAID 6 at all (RAID > 10 will withstand any two drive failure if you only use 4 drives), but > that's the reasoning. I think the best way to deal with the read-modify > write problem for RAID 6 is to use a small chunk size and deal with NxN > chunks as a unit. But YMMV. RAID10 will _not_ withstand any two-drive fail in a 4-drive scenario. If D1 and D3 fail, you're fscked D1 D2 D3 D4 True, I think that the point is that of the 5 possible 2 disk failures, 2 of them (in striped mirrors, not mirrored stripes) kill the array. For RAID5, all of them kill the array. But the fancy RAID setups are for _large_ arrays, not 4 disks, unless you are after the small write speed improvement (as I am). Plus any raid metadevice made of metadevices cannot autostart, which means tinkering during startup, which is only worth it for those large drive arrays. r --- Personalities : [raid0] [raid1] read_ahead 1024 sectors md0 : active raid0 md4[3] md3[2] md2[1] md1[0] 34517312 blocks 64k chunks