From owner-freebsd-isp Fri Oct 1 14:15:21 1999 Delivered-To: freebsd-isp@freebsd.org Received: from mercury.jorsm.com (mercury.jorsm.com [207.112.128.9]) by hub.freebsd.org (Postfix) with ESMTP id 2540B15280 for ; Fri, 1 Oct 1999 14:15:09 -0700 (PDT) (envelope-from jeff@mercury.jorsm.com) Received: from localhost (jeff@localhost) by mercury.jorsm.com (8.8.7/8.8.7) with SMTP id QAA29633; Fri, 1 Oct 1999 16:15:06 -0500 (CDT) Date: Fri, 1 Oct 1999 16:15:06 -0500 (CDT) From: Jeff Lynch To: Alan Judge Cc: freebsd-isp@FreeBSD.ORG Subject: Re: NetApp NFS & FreeBSD In-Reply-To: <19991001084551.7DAC615208@hub.freebsd.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-isp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Fri, 1 Oct 1999, Alan Judge wrote: > Jeff> Wow, 80Mbps (assuming b=bit, B=byte convention). > > Yep. I did a real quick test with a perl script creating 50000 > different 20K files. Results below, for a speed of around 3700KB/s or > 29Mb/s, so a good deal slow. But NFS latency issues may be involved > and using multiple processes could be faster. > > perl ~judgea/f.pl 2.69s user 22.45s system 9% cpu 4:25.60 total > > Watching with netstat -i, I/O drops off within a few seconds of the > end of the run, so there is little caching happening (on the FreeBSD > side). > > I'm not sure exactly what you wanted to test with caching, but I tried > writing 50000 different things to the same file, and got: > > perl ~judgea/f2.pl 1.77s user 11.78s system 6% cpu 3:24.06 total > > (or 4800KB/s, 38Mb/s) > > The difference is probably mostly directory and path handling stuff > and different caching behaviour on the Netapp, I'd guess. > > These are all write speed tests. > > I don't have time today to write a better benchmark. If you or anyone > wants to send me a script, I can probably run it. > -- > Alan Alan, thanks for your contribution. I would like to add a comparison benchmark on a Dual PII-400 with local 7500 baracudas on UW host adaptor. If you send me your scripts we'll run the same tests on a local (non RAID) file system. It will, of course be faster, but I want to quantify this for a decision process to go with a similar FBSD/Netapp setup. Perhaps others will volunteer to run the same tests on various FBSD h/w and s/w RAID systems. Thanks! --jeff ============================================================================ Jeffrey A. Lynch | JORSM Internet, Regional Internet Services email: jeff@jorsm.com | 7 Area Codes in Chicagoland and NW Indiana Voice: (219)322-2180 | 100Mbps+ Connectivity, 56K-DS3, V.90, ISDN Autoresponse: info@jorsm.com | Quality Service, Affordable Prices http://www.jorsm.com | Serving Gov, Biz, Indivds Since 1995 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-isp" in the body of the message