Date: Thu, 28 May 2009 12:16:52 +0300 From: Nikolay Denev <ndenev@gmail.com> To: Kip Macy <kmacy@freebsd.org> Cc: freebsd-fs@freebsd.org Subject: Re: RFT: ZFS MFC Message-ID: <3D66341E-3997-49A9-864A-1349B9B960D5@gmail.com> In-Reply-To: <3c1674c90905262138p7b196bc7ga2febb4667778e22@mail.gmail.com> References: <3c1674c90905151628h183cb1c2t8941843f8a828d4f@mail.gmail.com> <4A1C1E15.3080300@free.de> <3c1674c90905262138p7b196bc7ga2febb4667778e22@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --Apple-Mail-77-1054523524 Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Content-Transfer-Encoding: 7bit On May 27, 2009, at 7:38 AM, Kip Macy wrote: > On Tue, May 26, 2009 at 9:51 AM, Kai Gallasch <gallasch@free.de> > wrote: >> Kip Macy wrote: >>> I've MFC'd ZFS v13 to RELENG_7 in a work branch. Please test if >>> you can. >>> >>> http://svn.freebsd.org/base/user/kmacy/ZFS_MFC/ >>> >>> The standard disclaimers apply. This has only been lightly tested >>> in a >>> VM. Please do not use it with data you care about at this time. >>> >> >> Hi. >> >> http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/cddl/contrib/opensolaris/common/zfs/zfs_prop.c >> >> So ZFS v13 is now part of RELENG_7 MAIN. >> Did you receive any negative feedback until now? >> >> zpool upgrade: >> Should a ZFS v6 pool be upgraded through 'zpool upgrade' while in >> use? >> >> loader.conf: >> My old loader.conf (7.2-STABLE-amd64) settings for zfs: >> >> vm.kmem_size="3072M" >> vm.kmem_size_max="3072M" >> vfs.zfs.arc_min="120795878" >> vfs.zfs.arc_max="2899101078" >> # more stability? >> vfs.zfs.prefetch_disable=1 >> >> Is there a need to keep the old (painfully found :) known to work >> settings when upgrading from ZFS v6 to v13 on RELENG_7, because in >> your >> commit message I read: "- the arc now experiences backpressure from >> the >> vm (which can be too much - but this allows ZFS to work without any >> tunables on amd64)" >> > > Machines with less than 4GB, especially those with workloads with poor > locality, should have prefetch disabled. You should not need the > other tunables on amd64 (i386 is both a lower priority and a harder > problem). When reporting problems report them with the tunables > removed as well as whatever your existing tunables are. > > > -Kip How about just tuning down zfetch a bit? I'm currently using these settings : vfs.zfs.zfetch.array_rd_sz=65536 vfs.zfs.zfetch.block_cap=8 vfs.zfs.zfetch.max_streams=2 With them my machine seems to work better than with prefetch totally disabled, and does not do excessive IO like it does with the default prefetch settings while for example downloading/seeding a torrent file. But does this reduce memory usage too? Regards, Niki Denev --Apple-Mail-77-1054523524 content-type: application/pgp-signature; x-mac-type=70674453; name=PGP.sig content-description: This is a digitally signed message part content-disposition: inline; filename=PGP.sig content-transfer-encoding: 7bit -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.11 (Darwin) iEYEARECAAYFAkoeVoQACgkQHNAJ/fLbfrk+MwCgqXDr10IzCbvzZypIu0JS+1ex x3AAoK4WX7mD8krnxqura471/IATiCYA =CuBj -----END PGP SIGNATURE----- --Apple-Mail-77-1054523524--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3D66341E-3997-49A9-864A-1349B9B960D5>