From owner-freebsd-fs@FreeBSD.ORG Fri Feb 15 10:08:41 2013 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by hub.freebsd.org (Postfix) with ESMTP id 9072C5F0 for ; Fri, 15 Feb 2013 10:08:41 +0000 (UTC) (envelope-from grarpamp@gmail.com) Received: from mail-ob0-f179.google.com (mail-ob0-f179.google.com [209.85.214.179]) by mx1.freebsd.org (Postfix) with ESMTP id 63494E74 for ; Fri, 15 Feb 2013 10:08:41 +0000 (UTC) Received: by mail-ob0-f179.google.com with SMTP id un3so3371177obb.24 for ; Fri, 15 Feb 2013 02:08:35 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:x-received:in-reply-to:references:date:message-id :subject:from:to:content-type; bh=roeYBqcXmP9U9s6fy2Xa5Aj34u9VXWq7UKT7Pud/cYg=; b=NsGbMi/J3BjafbEx6xolVuICAy50VFafB10r38HV5j6elFx+o+e065BcVmHaw8pcRh RH+/U3tWhi/iGrN0FxL7H634YiXYcGmwYf4FG+nrCENCHVqfkJU0ELaWR0rs1KCWp2o9 pwIPPBhQot2of4dwDCYL7j23Btpga+IiyDGyJxNInCHOBSpDl+Jw2GU0wqlKnffuyHcE gic7eVGFfAS+MyTXww1SqpKBqryxLdZWzKeRX6RK+L7iK6geM9w8kgErhDxSCipStMHB PqK24CUub115vd7CmCQBJ/axv4wzobavfP6q4EIkyisRyt23x1FTUP3MdzqOSjawMB63 wuPg== MIME-Version: 1.0 X-Received: by 10.60.30.38 with SMTP id p6mr1400882oeh.2.1360922915348; Fri, 15 Feb 2013 02:08:35 -0800 (PST) Received: by 10.60.146.203 with HTTP; Fri, 15 Feb 2013 02:08:35 -0800 (PST) In-Reply-To: References: Date: Fri, 15 Feb 2013 05:08:35 -0500 Message-ID: Subject: Re: Crazy ZFS ZIL options: md(4) umass(4) From: grarpamp To: freebsd-fs@freebsd.org Content-Type: text/plain; charset=UTF-8 X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 15 Feb 2013 10:08:41 -0000 > ZIL is there for you to recover transactions in case of a crash. > It is your safety net. I always thought the ZIL was pushed out safely. So that still no matter what the disk would be consist [1]. Like when crash you just lose the ZIL's since the last ZIL push. Which odds are will be just work product here. > Use the RAM for ARC, it will provide more performance. But about reducing fragmentation without separate ZIL. I'm admittedly over full and will need to move data to new pool anyway. Just that with ZIL in main pool what article I read says problem can mostly come back without separate zil. I tend to run full till annoyed to redesign, bad habit. > If you don't have low latency SSD, put it on separate spinning disk This may be best. But ties to [1] above.. if ZIL spindle (even mirrored) dies, what is the point of what ZIL is on if ZIL separate from main pool is unsafe anyways. > USB ... unreliable I would have to test USB bus and devs for stablility. But I do have local lifetime warranty on 32GiB devices :) So maybe mirror 2 of them, or 2x2. Due to crypto I only get 7-15MiB/s on spindle anyway. > Why don't just set sync=disabled? Hmm, that might give performance as on any FS, but unsure about lesser fragmentation. I should find something to read on that. > What is split ZIL? Separate from main pool (spindles), 'zpool ... log ...' This may become an experiment :)