Date: Tue, 28 Aug 2007 12:54:35 +0200 From: Pawel Jakub Dawidek <pjd@FreeBSD.org> To: Simun Mikecin <numisemis@yahoo.com> Cc: freebsd-fs@freebsd.org Subject: Re: zfs and BIO_FLUSH on amr(4) Message-ID: <20070828105435.GC36596@garage.freebsd.pl> In-Reply-To: <937640.78988.qm@web36613.mail.mud.yahoo.com> References: <937640.78988.qm@web36613.mail.mud.yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--vEao7xgI/oilGqZ+ Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Tue, Aug 28, 2007 at 03:00:44AM -0700, Simun Mikecin wrote: > Using storage on amr(4) controller with zfs revealed reduced write perfor= mance. > It seems that writing to zfs concludes with a BIO_FLUSH call to flush the= controller cache to > disk. >=20 > amr(4) can have onboard cache memory with or without battery backup. Is B= IO_FLUSH call really > needed when amr(4) is used with onboard cache memory with a battery backu= p? I don't think so. The thing is that when ZFS receives information that write is done, it should be on disk (at some point) and you battery-backed cache should ensure that. You can turn off sending BIO_FLUSH by setting vfs.zfs.cache_flush_disable to 1 (in /boot/loader.conf). BTW. How big performance drop do you see with BIO_FLUSH turned on? --=20 Pawel Jakub Dawidek http://www.wheel.pl pjd@FreeBSD.org http://www.FreeBSD.org FreeBSD committer Am I Evil? Yes, I Am! --vEao7xgI/oilGqZ+ Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.4 (FreeBSD) iD8DBQFG0/7rForvXbEpPzQRAoiqAKChXdHzJ+RneGfiy816ePHBHAiPTQCfQuLp pI3ni9NuGwxkL2/8XdYyotQ= =GcIO -----END PGP SIGNATURE----- --vEao7xgI/oilGqZ+--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070828105435.GC36596>