From owner-freebsd-questions@FreeBSD.ORG Wed Feb 19 04:21:26 2014 Return-Path: Delivered-To: questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [8.8.178.115]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id D5EF697E; Wed, 19 Feb 2014 04:21:26 +0000 (UTC) Received: from mx1.fisglobal.com (mx1.fisglobal.com [199.200.24.190]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.freebsd.org (Postfix) with ESMTPS id 9A8DA1503; Wed, 19 Feb 2014 04:21:26 +0000 (UTC) Received: from smarthost.fisglobal.com ([10.132.206.191]) by ltcfislmsgpa02.fnfis.com (8.14.5/8.14.5) with ESMTP id s1J4KcqS021561 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NOT); Tue, 18 Feb 2014 22:20:38 -0600 Received: from THEMADHATTER (10.242.181.54) by smarthost.fisglobal.com (10.132.206.191) with Microsoft SMTP Server id 14.3.174.1; Tue, 18 Feb 2014 22:20:36 -0600 From: Sender: Devin Teske To: , References: In-Reply-To: Subject: RE: ZFS on Hardware RAID controller Date: Tue, 18 Feb 2014 20:20:01 -0800 Message-ID: <0fdd01cf2d29$dca0aba0$95e202e0$@FreeBSD.org> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Mailer: Microsoft Outlook 15.0 Thread-Index: AQLZ7+7/ieFjnRxrw7xfyVSgmKX96wLfKSMrAmxEjEMCYeNcAwLaiwfumFIvPAA= Content-Language: en-us X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10432:5.11.87, 1.0.14, 0.0.0000 definitions=2014-02-19_01:2014-02-18,2014-02-19,1970-01-01 signatures=0 Cc: dteske@freebsd.org X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.17 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 19 Feb 2014 04:21:26 -0000 > -----Original Message----- > From: freebsd@fongaboo.com [mailto:freebsd@fongaboo.com] > Sent: Tuesday, February 18, 2014 7:58 PM > To: questions@freebsd.org > Subject: Re: ZFS on Hardware RAID controller > > > I was speaking to someone else about this today, and it eventually became > apparent that we were getting into a sort-of Abbott-and-Costello-Who's-on-First > confusion... because apparently people can mean different things when they > use the term 'JBOD'. > > What I have always meant when I say 'JBOD' is a (not really) RAID mode that > simply concatenates the drives into one volume in a serial fashion, ie. > 'spanning'. Most RAID controllers and RAID-enabled NAS units that I have > interacted with in my life have offered this mode and referred to it as 'JBOD'. > This is not entirely correct. JBOD and RAID-SPAN are two different things. Your controller either supports one, or both (aside a host of other options, such as RAID-1, RAID-10; and if often RAID-5, maybe RAID-6). RAID-SPAN is RAID-0. Which is as you describe "simply concatenates the drives into one volume in a serial fashion, ie. 'spanning'." JBOD on the other hand stands for "Just a Bunch of Disks" and is not the same as "concatenating the disks" but rather allows your controller to throw the drives at the Operating System with the attitude of "here, you do it then." We can actually therefore call ZFS's RAID capabilities, "software RAID". However, it's a very _good_ software RAID that breaks the old adage that says "hardware RAID is faster than software RAID." To highlight the difference of "JBOD" versus "spanning" (aka RAID-0; aka RAID- SPAN), take the following use-case example: 1. You take 12 drives and apply spanning logic on the controller 2. You take a controller and put it into JBOD mode In scenario one (1), your OS still sees a single drive. In scenario two (2), your OS sees all twelve drives. Just wanted to clarify that putting a controller into JBOD-mode is not to be confused with RAID-0 or spanning (which would still be utilizing the controller's software; the best benefit of ZFS comes from letting it have direct access to each/every disk). > In this kind of mode, the motherboard and the OS still thinks it sees only a > single volume. So now I am gathering that this is also not ideal for ZFS, since it > would still not be aware of multiple physical volumes and be unable to optimize > accordingly. > > I'm learning for the first time that sometimes 'JBOD' can also refer to each > individual drive being mounted separately at least as far as the controller and > the motherboard is concerned. > > I just want to confirm 100% that this is how you are recommending multiple > drives be configured for ZFS. Because when I started the thread I was thinking > of JBOD as 'spanning'. > Correct; JBOD never means span (which is RAID-0). Rather, JBOD usually means using non-RAID capable hardware with a RAID-capable operating system (e.g., using ZFS). We are again in-essence talking about using software RAID to create a software pool. I don't want to go drawing similarities to many other software RAID solutions, because ZFS is truly in a class its own. But if you're familiar with the concepts of setting up a software RAID (mdadm, vinum, graid*, etc.) then ZFS should be more familiar. That's not to say that you need to know these things to use ZFS (in practice, ZFS has predictable commands with predictable syntax) but if you have ever created a software RAID you will be in unique position to better understand the JBOD mentality. So in transitioning to a test platform that uses "JBOD" you essentially have two choices that look attractive... + Ditch the RAID card and use a standard adapter for connecting your drives + Find a way to change your RAID controller to export all the disks NB: This may involve using a LSI provided utility or flashing a QLogic card NB: If you give us your exact card info, someone may have information on how to transition it into JBOD mode for ZFS + Create a bunch of single-disk RAID-0 arrays (12 disks? 12x single-disk SPAN or RAID-0 arrays; producing 12x single LUNs for use in ZFS). Hope this gives some ideas. -- Devin > On Tue, 18 Feb 2014, Eitan Adler wrote: > > > On Tue, Feb 18, 2014 at 5:27 PM, wrote: > >> > >> When we spoke, you noted that when installing ZFS on multiple disks > >> connected to a hardware RAID controller, it is best to config it to JBOD. > > > > There are a few reasons for this. > > (a) Hardware RAID serves as a single point of failure: if the > > contoller dies you have neither disk > > (b) As Andrew noted , using hardware RAID means that ZFS won't be able > > to tell which disk is which. The ZFS management tools won't work as > > expected (they will show only one disk). > > (c) Since ZFS implements RAID itself it can use knowledge about the > > physical disks for better performance > > > > Also see: https://en.wikipedia.org/wiki/ZFS#ZFS_and_hardware_RAID > > > >> I tried to explain this to a colleague, but they were skeptical. > >> Would you (or anyone) be willing to give me a one or two line > >> sales-pitch on > > > > "ZFS does RAID better than the controller." > > > >> why one > >> should abandon traditional notions of RAID performance in favor of > >> allowing ZFS to do disk management? > > > > The goal isn't to give up on RAID but move its implementation to ZFS. > > > > -- > > Eitan Adler > > Source, Ports, Doc committer > > Bugmeister, Ports Security teams > > _______________________________________________ > > freebsd-questions@freebsd.org mailing list > > http://lists.freebsd.org/mailman/listinfo/freebsd-questions > > To unsubscribe, send any mail to "freebsd-questions- > unsubscribe@freebsd.org" > > > _______________________________________________ > freebsd-questions@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-questions > To unsubscribe, send any mail to "freebsd-questions-unsubscribe@freebsd.org" _____________ The information contained in this message is proprietary and/or confidential. If you are not the intended recipient, please: (i) delete the message and all copies; (ii) do not disclose, distribute or use the message in any manner; and (iii) notify the sender immediately. In addition, please be aware that any message addressed to our domain is subject to archiving and review by persons other than the intended recipient. Thank you.