From owner-freebsd-fs@FreeBSD.ORG Wed Aug 11 22:10:14 2010 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 7BFEA106566B for ; Wed, 11 Aug 2010 22:10:14 +0000 (UTC) (envelope-from artemb@gmail.com) Received: from mail-vw0-f54.google.com (mail-vw0-f54.google.com [209.85.212.54]) by mx1.freebsd.org (Postfix) with ESMTP id 27CBC8FC15 for ; Wed, 11 Aug 2010 22:10:13 +0000 (UTC) Received: by vws7 with SMTP id 7so577938vws.13 for ; Wed, 11 Aug 2010 15:10:13 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:sender:received :in-reply-to:references:date:x-google-sender-auth:message-id:subject :from:to:content-type:content-transfer-encoding; bh=2e7d9d46dA8nK1FDd35uuaxgai7N5E+JMKq+VmOlybQ=; b=QytbLFni/S24vpmqzRuOukSjbCJcYF8V0kd05/h/X/wd0ZkC617hn8Z/RrU8i995nW /i0T8dAJCASyd5oN7UvAifxYwnVTaQQbrf20TAgX7bhBLWkNYn51rVQhD6WS8CHYYCV5 V8MMJcdF06eMXHQCMtiso7hmOPbIXRubw4Rho= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:in-reply-to:references:date :x-google-sender-auth:message-id:subject:from:to:content-type :content-transfer-encoding; b=EM92v6zoRI+aEpu8w8xc+eSZ0AULKLR3bcLD5tmAOlcF7v9LUZcElyI5Xt2ri+wFAh To3/fh3UJlp9Zzu3/AbFEiyasI3sWLpEiAd8FStcxJXl9bUObnspMoEC0HQqv5KVGBwC FaUwRlGL5Ciub5fMkgS0UmRL9q4KpgnzcjVTo= MIME-Version: 1.0 Received: by 10.220.96.4 with SMTP id f4mr11885756vcn.267.1281564613074; Wed, 11 Aug 2010 15:10:13 -0700 (PDT) Sender: artemb@gmail.com Received: by 10.220.49.70 with HTTP; Wed, 11 Aug 2010 15:10:12 -0700 (PDT) In-Reply-To: <20100811214302.GB44635@tolstoy.tols.org> References: <20100810214418.GA28288@tolstoy.tols.org> <20100811014919.GA52992@icarus.home.lan> <20100811192537.GA44635@tolstoy.tols.org> <20100811214302.GB44635@tolstoy.tols.org> Date: Wed, 11 Aug 2010 15:10:12 -0700 X-Google-Sender-Auth: IQW6hoAIXNctRbAKEZRPrxBpB_0 Message-ID: From: Artem Belevich To: freebsd-fs@freebsd.org Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Subject: Re: zfs arc - just take it all and be good to me X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 11 Aug 2010 22:10:14 -0000 On Wed, Aug 11, 2010 at 2:43 PM, Marco van Tol wrote: >> There's a hack floating around that attempts to force kernel into >> freeing up memory from inactive/cache lists before draining ARC. It >> does help a bit with this issue, but it's still a hack. > > That makes sense Artem, thanks. =A0I think you mean the posts with the > perl one-liner I used in my tests as well. =A0(perl variable assignment o= f > 1.5GB in the posts their version) Well, that perl one-liner is somewhat similar to using guillotine to cure a minor headache. I was actually referring to the patch mentioned in this post: http://old.nabble.com/Re%3A-Serious-zfs-slowdown-when-mixed-with-another-fi= le-system-(ufs-msdosfs-etc.).-p29137467.html The patch itself is here: http://pastebin.com/ZCkzkWcs > I had seen the posts that mentioned that one and decided to remember the > perl hack. :) > > What I understand from it: > - In a UFS/ZFS mixed system > - In a scenario where UFS "page cache" took (almost) all available memory > - Run a perl one-liner to throw out the UFS active/inactive usage > - Kind-off hope you do enough relevant ZFS accesses that you get a good > =A0new situation. > > So, if my worries can shift from fighting with kmem_size and arc_max to > fighting with arc_min, that's a fight I like a lot better. =A0Especially > on zfs-only systems, I have to admit. Non-ZFS filesystems are not the only entities that cause memory to end up on inactive list. It may end up there for number of other reasons. For instance on my ZFS-only box I currently have ~1G worth of it. Filesystem just happens to be the entity that often stashes a lot of data on inactive list and that causes immediate issues for ZFS. --Artem