From owner-freebsd-fs Mon Oct 5 11:51:09 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id LAA01844 for freebsd-fs-outgoing; Mon, 5 Oct 1998 11:51:09 -0700 (PDT) (envelope-from owner-freebsd-fs@FreeBSD.ORG) Received: from smtp02.primenet.com (smtp02.primenet.com [206.165.6.132]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id LAA01779; Mon, 5 Oct 1998 11:50:47 -0700 (PDT) (envelope-from tlambert@usr01.primenet.com) Received: (from daemon@localhost) by smtp02.primenet.com (8.8.8/8.8.8) id LAA17982; Mon, 5 Oct 1998 11:50:14 -0700 (MST) Received: from usr01.primenet.com(206.165.6.201) via SMTP by smtp02.primenet.com, id smtpd017960; Mon Oct 5 11:50:08 1998 Received: (from tlambert@localhost) by usr01.primenet.com (8.8.5/8.8.5) id LAA23549; Mon, 5 Oct 1998 11:50:05 -0700 (MST) From: Terry Lambert Message-Id: <199810051850.LAA23549@usr01.primenet.com> Subject: Re: filesystem safety and SCSI disk write caching To: oppermann@pipeline.ch (Andre Oppermann) Date: Mon, 5 Oct 1998 18:50:04 +0000 (GMT) Cc: gpalmer@FreeBSD.ORG, freebsd-fs@FreeBSD.ORG In-Reply-To: <3618DA15.46329AA1@pipeline.ch> from "Andre Oppermann" at Oct 5, 98 04:39:17 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-fs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > > I can post (once again) the results of a Novell study on server usage > > > > patterns. The 30,000 foot view for a typical server breaks down to: > > > > > > > > 75% reads > > > > 15% writes > > > > 8% directory search operations > > > > 2% other > > > > I think that is very dependant on the server type. PC NetWare fileservers > > probably have very different access patterns to (say) a web server or a mail > > server. Let alone a news server. > > Is there a way to gather such statistics on FreeBSD? > > I'd like to run it on all my boxes (and others) to get representative > figures. After that we can discuss optimizations. This would be a very good idea. It would be best to instrument at the lowest level, *below* where soft updates does its implicit write gathering, and/or any elevator sorting occurs. I have suggested one set of instrumentation, which is interval sampling of number of tagged commands outstanding, which would resolve once and for all the current debate. Unfortunately, I don't have the good firmware that Justin seems to have, and thus I can't really make the measurements myself. It also seemed to me that the authors of the CAM code would be in a better position about where to validly determine max tag loading than I, and so would be better suited to the task of coding up the statistics gathering. One method of testing FreeBSD, but which would include testing SAMBA, would be to run the Ziff Davis "NetBench" suite against a FreeBSD machine acting as a server. There are similar benchmarks (LANBench?) available for download from Novell, laast time I looked at the FTP site. 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