From owner-freebsd-questions@FreeBSD.ORG Tue Aug 10 17:24:41 2010 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 128A91065679 for ; Tue, 10 Aug 2010 17:24:41 +0000 (UTC) (envelope-from gull@gull.us) Received: from mail8.sea5.speakeasy.net (mail8.sea5.speakeasy.net [69.17.117.53]) by mx1.freebsd.org (Postfix) with ESMTP id E05EC8FC17 for ; Tue, 10 Aug 2010 17:24:40 +0000 (UTC) Received: (qmail 24063 invoked from network); 10 Aug 2010 17:24:40 -0000 Received: from dsl081-163-112.sea1.dsl.speakeasy.net (HELO ringbill.gull.us) ([64.81.163.112]) (envelope-sender ) by mail8.sea5.speakeasy.net (qmail-ldap-1.03) with AES256-SHA encrypted SMTP for ; 10 Aug 2010 17:24:40 -0000 Received: from alphonse ([192.168.10.9] helo=alphonse.gull.us) by ringbill.gull.us with esmtpa (Exim 4.71 (FreeBSD)) (envelope-from ) id 1OisZ5-0002XD-RO for freebsd-questions@freebsd.org; Tue, 10 Aug 2010 10:24:39 -0700 Message-Id: <998D8501-AD49-4436-9D5C-E0C3A9ECC872@gull.us> From: David Brodbeck To: FreeBSD Questions In-Reply-To: Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Content-Transfer-Encoding: 7bit Mime-Version: 1.0 (Apple Message framework v936) Date: Tue, 10 Aug 2010 10:24:39 -0700 References: X-Mailer: Apple Mail (2.936) Subject: Re: ZFS practical application? X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 10 Aug 2010 17:24:41 -0000 On Aug 9, 2010, at 3:40 PM, Ed Flecko wrote: > Hi folks, > I've been reading about the ZFS file system, and I'm having a hard > time understanding maybe the most practical business application(s)? > > I think I understand a little bit about it (from a conceptual > perspective) that it's a self-healing 128 bit filesystem, better data > integrity checking, etc. > > I have a small business (< 50 end users) and I'm wondering perhaps > some examples that you might think would be most applicable for a > FreeBSD server(s) and the ZFS filesystem > > One of the things that seems like might be a detriment as well as an > asset, is it's ability to expand as necessary, but then I'm wondering > what prevents the filesystem from just "running away"? You can set a quota for each filesystem that it won't grow beyond. You can also set reservations to ensure a given filesystem will get a certain amount of space, even if other filesystems grow. With intelligent use of these features you don't have to worry much about "runaway" filesystems. ZFS is very handy for situations where you have a large storage pool that you want to split up for different users and applications. It's much more flexible than a rigid partitioning scheme; you can safely and quickly resize filesystems to best use the available space. I've also found the compression feature to be quite effective on filesystems that store data that compresses well. We have an NFS share that stores mainly text, and with the default lzjb compression I've seen 1.5:1 ratios with no detectable performance hit. (Reads actually got slightly *faster*, but that may have been a testing glitch.) gzip compression achieved much higher compression ratios but started to affect performance. I expect even better results when we eventually deploy ZFS deduplication. ZFS snapshots are handy for recovering deleted user files without having to restore from backup. NB: We're currently running OpenSolaris on our fileservers but I'm going to look into switching to FreeBSD now that ZFS on FreeBSD is a bit more mature. I've gotten kind of disenchanted with OpenSolaris's slow update cycle.