From mboxrd@z Thu Jan 1 00:00:00 1970 From: Reindl Harald Subject: Re: proactive disk replacement Date: Tue, 21 Mar 2017 13:03:22 +0100 Message-ID: <8b108a89-9d17-63f0-de1c-80b17be4411a@thelounge.net> References: <3FA2E00F-B107-4F3C-A9D3-A10CA5F81EC0@allygray.2y.net> <11c21a22-4bbf-7b16-5e64-8932be768c68@websitemanagers.com.au> <02316742-3887-b811-3c77-aad29cda4077@websitemanagers.com.au> <583576ca-a76c-3901-c196-6083791533ee@thelounge.net> <20170321113447.GA18665@metamorpher.de> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20170321113447.GA18665@metamorpher.de> Sender: linux-raid-owner@vger.kernel.org To: Andreas Klauer Cc: Adam Goryachev , Jeff Allison , linux-raid@vger.kernel.org List-Id: linux-raid.ids Am 21.03.2017 um 12:34 schrieb Andreas Klauer: > On Tue, Mar 21, 2017 at 12:03:51PM +0100, Reindl Harald wrote: >> but the point is that with RAID5/6 the recovery itself is *heavy random >> IO* and that get *combined* with the random IO auf the normal workload >> and that means *heavy load on the disks* > > Where do you get that random I/O idea from? Rebuild is linear. > Or what do you mean by random I/O in this context? (RAID rebuilds) > What kind of random things do you think the RAID is doing? the IO of a RAID5/6 rebuild is hardly linear beause the informations (data + parity) are spread all over the disks while in case of RAID1/10 it is really linear > If you see read errors during rebuild, the most common cause is > that the rebuild also happens to be the first read test since forever. > (Happens to be the case for people who don't do any disk monitoring.) > >> here you go: http://www.zdnet.com/article/why-raid-6-stops-working-in-2019/ > > This is just wrong