From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?ISO-8859-1?Q?Kristleifur_Da=F0ason?= Subject: Re: 3 disk RAID1? Date: Mon, 8 Mar 2010 22:18:52 +0000 Message-ID: <73e903671003081418g13861e21pa412f46995ad9781@mail.gmail.com> References: <5bdc1c8b1003081239l32c5653fs8347179e23c0a287@mail.gmail.com> <20100308205803.GA31731@cthulhu.home.robinhill.me.uk> <5bdc1c8b1003081304n7a59e99ej1f66078090980f06@mail.gmail.com> <4877c76c1003081411p3c729c89q3d908d07f19b0857@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <4877c76c1003081411p3c729c89q3d908d07f19b0857@mail.gmail.com> Sender: linux-raid-owner@vger.kernel.org To: Mark Knecht Cc: Linux-RAID List-Id: linux-raid.ids On Mon, Mar 8, 2010 at 10:11 PM, Michael Evans = wrote: > On Mon, Mar 8, 2010 at 1:04 PM, Mark Knecht wr= ote: >> On Mon, Mar 8, 2010 at 12:58 PM, Robin Hill = wrote: >>> On Mon Mar 08, 2010 at 12:39:26PM -0800, Mark Knecht wrote: >>> >>>> Hi all, >>>> =A0 =A0I'm still very much on a steep learning curve about what I = can do >>>> with Linux software RAID. In another thread this weekend a couple = of >>>> responders discussed among themselves 3-disk RAID1 solutions that = can >>>> survive if 2 disks die. I don't understand what that means. Can >>>> someone point me at a quick explanation? Is that really possible? >>>> >>>> =A0 =A0In general I'm using a few Wikipedia pages and gravitate to= ward the >>>> diagrams as much as anything. >>>> >>>> http://en.wikipedia.org/wiki/RAID >>>> http://en.wikipedia.org/wiki/RAID1#RAID_1 >>>> >>>> RAID0 - striping, speed not reliability (2 disk minimum) >>>> RAID1 - duplicate data, no other protection (2 disk minimum) >>>> >>>> =A0 =A0How do I build RAID1 using three drives? Just duplicate the= data 3 >>>> times? If drives start going bad how do I determine which one or t= wo >>>> are failing? (fsck? SMART?) With 3 drives 1 fail seems relatively >>>> straightforward to figure out, but 2? >>>> >>> A 3-disk RAID1 is just 3 duplicate copies, yes. =A0And RAID only pr= otects >>> against hardware failures, so you know which disk has failed becaus= e it >>> gets kicked out of the array as faulty. =A0This is the same regardl= ess of >>> how many mirrored copies you have (md will detect a write failure t= o a >>> drive and mark it as faulty - read errors will cause the failed blo= ck to >>> get rewritten). >>> >>> As for how to create it - it's just the same process as for a 2-dis= k >>> RAID1 but specifying 3 drives (assuming you're using Linux md softw= are >>> RAID - if not, please specify what you're intending to use). =A0The= manual >>> page for mdadm should give you everything you need - do ask if ther= e's >>> anything you want clarifying though. >>> >>> Cheers, >>> =A0 =A0Robin >> >> Thanks Robin. Maybe I am getting smarter about this if I'm figuring >> out what others are talking about! ;-) >> >> Cheers, >> Mark >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-raid= " in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at =A0http://vger.kernel.org/majordomo-info.html >> > > When in doubt, read the manual two or three more times. > > This might also help you: http://wiki.tldp.org/LVM-on-RAID =A0I wrote > some background comparison sections when I made that... Good advice. The contents of the manual didn't really stick for me until I'd actually *done* the tasks I was trying to figure out. The beautiful and classic Catch-22 of learning new things. So I somewhat empathically suggest you obtain a virtual machine program and start up a few disposable virtual machines. (VirtualBox is pretty good and also free.) Try creating some RAID's on VM's and getting them to boot, try failing a device and readding it, etc. And keep the manual in a window next to the virtual machine's window :) -- Kristleifur -- To unsubscribe from this list: send the line "unsubscribe linux-raid" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html