Date: Tue, 27 Mar 2012 07:48:55 -0400 From: John Baldwin <jhb@freebsd.org> To: freebsd-fs@freebsd.org Subject: Re: NFSv3, ZFS, 10GE performance Message-ID: <201203270748.55392.jhb@freebsd.org> In-Reply-To: <20120326190342.0b78cbc8@fabiankeil.de> References: <4F703815.8070809@crashme.org> <alpine.GSO.2.01.1203261146200.22350@freddy.simplesystems.org> <20120326190342.0b78cbc8@fabiankeil.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On Monday, March 26, 2012 1:03:42 pm Fabian Keil wrote: > Bob Friesenhahn <bfriesen@simple.dallas.tx.us> wrote: > > > On Mon, 26 Mar 2012, Sven Brandenburg wrote: > > > > > > Hopefully, readahead doesn't kill performance for smaller files.. :-) > > > > You are right to be concerned. There are plenty of cases where > > read-ahead damages application performance. Reading data which is > > never actually used is expensive. > > > > It would be useful if FreeBSD would support posix_fadvise() so that > > applications can specify the type of access they will use, and if this > > advice can be used by NFS and the filesystem layer to decide if > > read-ahead should be used, and how much. > > posix_fadvise() is already available in FreeBSD 10.0-CURRENT. It doesn't quite do the trick of bumping up the read ahead amount for sequential yet (though that is an easy change). Implementing WILLNEED is a bit trickier as it requires per-FS support. I have patches to implement it for UFS. -- John Baldwin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201203270748.55392.jhb>