From owner-freebsd-fs@freebsd.org Fri Apr 29 15:03:48 2016 Return-Path: Delivered-To: freebsd-fs@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 65680B20647 for ; Fri, 29 Apr 2016 15:03:48 +0000 (UTC) (envelope-from eborisch@gmail.com) Received: from mail-oi0-x233.google.com (mail-oi0-x233.google.com [IPv6:2607:f8b0:4003:c06::233]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (Client CN "smtp.gmail.com", Issuer "Google Internet Authority G2" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 2B7321A17 for ; Fri, 29 Apr 2016 15:03:48 +0000 (UTC) (envelope-from eborisch@gmail.com) Received: by mail-oi0-x233.google.com with SMTP id k142so121418091oib.1 for ; Fri, 29 Apr 2016 08:03:48 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc; bh=tLhF5UbD5AOG7Sj7Ve7AkmtyYirMvvTBciH2WpLuw3c=; b=vrNqvolcUrrBgbdpMHCLdFbosvA0hmkk0lWAWA1JKAOQcIdRz/qPBuofm+RjcnfD+r umHs7z+ID20QuPV/y5RHZ5wcwey8n855XXq5GLfViL0EEdlCMU3lU2ohcNtogr8FOb6k rxwDUjLGiAFTYC/FPibBBfcsBSpbGZZ9pfguy7m3QmDJU6JWiRihXeIESEByd0hymj0k eUxmZiikVqkjN8zmQHxXV0Y6Iq8aR9YApwje+/vDQljh3+2pnrUELNYe/uQdVrns4PPj oz4Wi+O2k9xY0F0UB1C2XiEgUfAto6qmApoe4HaCIaKYuUWocf9RSVbBPj/v89JELWH6 vkOg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:mime-version:in-reply-to:references:date :message-id:subject:from:to:cc; bh=tLhF5UbD5AOG7Sj7Ve7AkmtyYirMvvTBciH2WpLuw3c=; b=fWOhJ6BHbuDH5xoCAdqhVg+SWnqEmX5YxGywrJTZxuZMceBaXJjjD9q+RatFR1qrcc VmLdvW3BndZcOl08ii788GSt3kbYwsf4m+n1su6Wi2Qz3OIoYn6lt5g9CsX8unBuQKAL LYzXM1b2+Wgu7dEz2iO4JgDteVzTwuXBt9GHjJwiPkXj8eW4IkEJFCj+GJbDtRdFT3NN 89a2ehBtWDrnC4XOdgFEp+PdzL2fNoLfrvtRFLEUX4di61agbTSB9MTTJfw9WXQ2/9r6 4mKfV8mXACYWSzOlpkIOqVfWQZvKScJlnNkuGuFnp62kRkidrVUbKdtmEOd0OV2iSMf5 keQg== X-Gm-Message-State: AOPr4FWfh/0/G9Qr19vOmEsKW3UfDrDQFgwZ0hYadTru+e4Cyuqo7Vch1wbK1JlMOzQ/g47btgWYMrG0zTzOjw== MIME-Version: 1.0 X-Received: by 10.202.189.10 with SMTP id n10mr8897147oif.101.1461942227278; Fri, 29 Apr 2016 08:03:47 -0700 (PDT) Received: by 10.60.52.145 with HTTP; Fri, 29 Apr 2016 08:03:47 -0700 (PDT) In-Reply-To: References: <381846248.2672053.1461695277122.JavaMail.yahoo.ref@mail.yahoo.com> <381846248.2672053.1461695277122.JavaMail.yahoo@mail.yahoo.com> <1461736217.1121.17.camel@michaeleichorn.com> <08d59afe-c835-fa8d-0e52-78afcb1cc030@denninger.net> Date: Fri, 29 Apr 2016 10:03:47 -0500 Message-ID: Subject: Re: How to speed up slow zpool scrub? From: "Eric A. Borisch" To: Ronald Klop Cc: "freebsd-fs@freebsd.org" , Karl Denninger Content-Type: text/plain; charset=UTF-8 X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 29 Apr 2016 15:03:48 -0000 On Fri, Apr 29, 2016 at 9:28 AM, Ronald Klop wrote: > On Fri, 29 Apr 2016 15:47:22 +0200, Karl Denninger >> ZFS makes the *assumption*, fair or not, that everything in its >> RAM-based caches is correct. If that assumption is violated you will >> eventually be a very sad Panda. Use ECC memory or don't use ZFS. >> > > Just like UFS makes an assumption about correct memory and correct disks? > > ECC helps ZFS as much as ECC helps UFS. > And without ECC ZFS provides more failsafes than UFS. But nothing is > perfect. > You guys make it sound like ZFS has no added benefits if you don't use ECC, > which is not true. > > UFS < ZFS < ZFS+ECC > And UFS+ECC is somewhere in between probably. > As long as people understand the risks/benefits things are ok. I a key distinction is that UFS has fsck for attempting to repair inconsistencies on the drive, where ZFS does not have a similar tool, because "[t]he only way for inconsistent data to exist on disk in a ZFS configuration is through hardware failure [...] or when a bug exists in the ZFS software." [1] So if ZFS fails, it is more likely to fail hard; enter the ECC (avoid hardware failure) "requirement". I personally have one system running without ECC, but it is a tiny system at home that serves as the firewall the cable modem runs into. It is backed up and stores nothing of real value on the media, but I love having ZFS on it because I can do things like beadm for upgrades, or diffs of /etc files with previous (automated) snapshots. (It's also running with less than 4G of RAM, tsk-tsk...) If you are storing data you care about* on a ZFS system without ECC, you are doing it wrong. If that system *is* the backup, you are in a gray area depending on your risk profile. (Is it OK if the backups fail, because I still have the source and am willing to risk having only one copy while I rebuild the backup?) So I'd temper Karl's statement to "Use ECC memory -- or really understand the risks -- or don't use ZFS." - Eric [1] http://docs.oracle.com/cd/E19253-01/819-5461/6n7ht6r6p/index.html * can't afford to lose / can't recreate