Date: Thu, 14 Dec 2000 23:34:48 +0000 (GMT) From: Terry Lambert <tlambert@primenet.com> To: bright@wintelcom.net (Alfred Perlstein) Cc: tlambert@primenet.com (Terry Lambert), henrich@sigbus.com (Charles Henrich), freebsd-fs@FreeBSD.ORG Subject: Re: Filesystem tuning (minimize seeks) Message-ID: <200012142334.QAA15873@usr08.primenet.com> In-Reply-To: <20001214150353.J4589@fw.wintelcom.net> from "Alfred Perlstein" at Dec 14, 2000 03:03:54 PM
next in thread | previous in thread | raw e-mail | index | archive | help
> > The NFS protocol is defined as not returning success unless the > > write has been committed to stable storage. In FreeBSD, this > > tends to serialize NFS I/O from a single client, and between > > multiple clients in excess of the number of nfsiod's you are > > running. > > This is untrue for NFSv3, that's why there are write and commit > RPCs. By using write ahead then delaying the commit you can > increase performance by only stalling out a single nfsiod to > sync out a large section of a file. The sync still has to occur before the client returns to the user space program that the write has been successful. This works when you turn a big write into a bunch of little write at the mount write size. For multiple user space write operations, the client still stalls between the writes until after the commit. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-fs" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200012142334.QAA15873>