From owner-freebsd-stable Tue Mar 20 11: 0: 9 2001 Delivered-To: freebsd-stable@freebsd.org Received: from sdmail0.sd.bmarts.com (sdmail0.sd.bmarts.com [209.247.77.155]) by hub.freebsd.org (Postfix) with ESMTP id A4DD337B719 for ; Tue, 20 Mar 2001 11:00:06 -0800 (PST) (envelope-from gordont@bluemtn.net) Received: from localhost (gordont@localhost) by sdmail0.sd.bmarts.com (8.11.3/8.11.2/BMA1.1) with ESMTP id f2KIxLK12604; Tue, 20 Mar 2001 10:59:21 -0800 (PST) Date: Tue, 20 Mar 2001 10:59:21 -0800 (PST) From: Gordon Tetlow X-X-Sender: To: Alfred Perlstein Cc: Jan Conrad , Subject: Re: NFS performance In-Reply-To: <20010320105404.D29888@fw.wintelcom.net> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Tue, 20 Mar 2001, Alfred Perlstein wrote: > Actually, from what I've been told, TCP allows for much larger requests > than what UDP does, afaik UDP maxes out at 8k while tcp should be able > to go to 32k (maybe 64k) and give possibly better performance. This is true. I'm used to working with nfs transaction sizes of < 500 bytes. > Plus each time you 'hickup' under a UDP mount it's a lot more painful > because since it's possible for each RPC to be broken into several > packets you have a lot more retransmition to do. Yup. This is why a reliable (switched) LAN is needed. > However, as the manpages and sysadmin books say, NFS tuning is a > black art and results under various setting may vary wildly. It's very application specific which is why no single formula works. Reading up on the subject is definitely required. -gordon To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message