From mboxrd@z Thu Jan 1 00:00:00 1970 From: Roberto Spadim Subject: Re: [patch 2/3 v3] raid1: read balance chooses idlest disk for SSD Date: Mon, 2 Jul 2012 01:36:53 -0300 Message-ID: References: <20120702010840.197370335@kernel.org> <20120702011031.890864816@kernel.org> <20120702030245.GB29770@kernel.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: Sender: linux-raid-owner@vger.kernel.org To: Shaohua Li Cc: linux-raid@vger.kernel.org, neilb@suse.de, axboe@kernel.dk List-Id: linux-raid.ids sorry about many posts, i will stop a bit now... could we turn off disk NCQ? ok nobody want to read this, but could we turn off NCQ and get better performace? since we do a good elevator at linux code... maybe turn off disk cache too... (it=B4s only 32MB for sata/sas, and can be alot to raid controller...), just some ideas to get better performace... (i didn=B4t think before write if cache help o= r not when we send I/O to disk, but check if it=B4s a good question or not) 2012/7/2 Roberto Spadim : > for SSD without inteligent parts (cache and queue) i think this could > works nice... (at least in theory) > for USB sticks and microsd cards may work too... > > > 2012/7/2 Shaohua Li : >> On Sun, Jul 01, 2012 at 11:13:42PM -0300, Roberto Spadim wrote: >>> nice =3D) very very nice =3D) >>> maybe to get better than this... select the disk with min(pending t= ime) >>> >>> time could be estimated with something like: >>> (distance * time/distance unit) + (blocks to read/write * time to >>> read/write 1 block) + (non sequencial penalty time) >> >> I didn't think there is way to measure request time. Disks support N= CQ, they >> can dispatch several requests at one time and finish them almost at = the same >> time. I used to measure this time (to make CFQ ioscheduler self tune= ), but >> failed. >> >> Thanks, >> Shaohua >> -- >> 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 http://vger.kernel.org/majordomo-info.html > > > > -- > Roberto Spadim > Spadim Technology / SPAEmpresarial --=20 Roberto Spadim Spadim Technology / SPAEmpresarial -- 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