From owner-freebsd-stable@freebsd.org Tue Feb 9 17:05:25 2016 Return-Path: Delivered-To: freebsd-stable@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 430BAAA362D for ; Tue, 9 Feb 2016 17:05:25 +0000 (UTC) (envelope-from hausen@punkt.de) Received: from kagate.punkt.de (kagate.punkt.de [217.29.33.131]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id D89E3C51; Tue, 9 Feb 2016 17:05:24 +0000 (UTC) (envelope-from hausen@punkt.de) Received: from hugo10.ka.punkt.de (hugo10.ka.punkt.de [217.29.44.10]) by gate2.intern.punkt.de with ESMTP id u19H5Mcc069405; Tue, 9 Feb 2016 18:05:22 +0100 (CET) Received: from [217.29.44.208] ([217.29.44.208]) by hugo10.ka.punkt.de (8.14.2/8.14.2) with ESMTP id u19H5Mso078168; Tue, 9 Feb 2016 18:05:22 +0100 (CET) (envelope-from hausen@punkt.de) Content-Type: text/plain; charset=utf-8 Mime-Version: 1.0 (Mac OS X Mail 9.2 \(3112\)) Subject: Re: Best practices for ZFS setup for a strictly SSD based system? From: "Patrick M. Hausen" In-Reply-To: Date: Tue, 9 Feb 2016 18:05:25 +0100 Cc: freebsd-stable Content-Transfer-Encoding: quoted-printable Message-Id: References: <2D296837-3B06-4E72-B8B0-A33AE6CE48AE@punkt.de> To: Alan Somers X-Mailer: Apple Mail (2.3112) X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 09 Feb 2016 17:05:25 -0000 Hi! > Am 09.02.2016 um 17:32 schrieb Alan Somers : > [...] > http://blog.delphix.com/matt/2014/06/06/zfs-stripe-width/ >=20 >>=20 >> * Will e.g. MySQL still profit from residing on a mirror >> instead of a RAIDZ2, even if all disks are SSDs? >=20 > Yes, because a mirrored vdev has as many read IOPs as all of its disks > combined. So a RAID10 of SSDs will have many read IOPs indeed. Ah =E2=80=A6 yes. Now I remember :) > [=E2=80=A6] > Will MySQL access its files in fixed-size records? If so, you can set > the recsize filesystem property accordingly. If not, you should > probably leave recsize at the default. If you profile MySQL's disk > accesses and determine that there is a dominant recordsize, then go > ahead and set ZFS's recsize to the next highest power of two. >=20 > As usual, disable atime. We already knew these. But thanks a lot for the vdev setup hints! So it will be a mirror for OS and DB and a 4+2 raidz2 for the rest of the data. Our MySQL zvols are currently set up like this: DB files: recordsize=3D16k atime=3Doff primarycache=3Dmetadata InnoDB log files: recordsize=3D128k (rest inherited from above) Kind regards, Patrick --=20 punkt.de GmbH * Kaiserallee 13a * 76133 Karlsruhe Tel. 0721 9109 0 * Fax 0721 9109 100 info@punkt.de http://www.punkt.de Gf: J=C3=BCrgen Egeling AG Mannheim 108285