From owner-freebsd-questions@freebsd.org Wed Jun 21 16:28:01 2017 Return-Path: Delivered-To: freebsd-questions@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 02810D93654 for ; Wed, 21 Jun 2017 16:28:01 +0000 (UTC) (envelope-from galtsev@kicp.uchicago.edu) Received: from cosmo.uchicago.edu (cosmo.uchicago.edu [128.135.20.71]) by mx1.freebsd.org (Postfix) with ESMTP id D5F9272F05 for ; Wed, 21 Jun 2017 16:28:00 +0000 (UTC) (envelope-from galtsev@kicp.uchicago.edu) Received: by cosmo.uchicago.edu (Postfix, from userid 48) id A7A51CB8CE1; Wed, 21 Jun 2017 11:27:53 -0500 (CDT) Received: from 128.135.52.6 (SquirrelMail authenticated user valeri) by cosmo.uchicago.edu with HTTP; Wed, 21 Jun 2017 11:27:53 -0500 (CDT) Message-ID: <41839.128.135.52.6.1498062473.squirrel@cosmo.uchicago.edu> In-Reply-To: <45657887-638e-bb6d-c318-7046fdea1ca6@tundraware.com> References: <45657887-638e-bb6d-c318-7046fdea1ca6@tundraware.com> Date: Wed, 21 Jun 2017 11:27:53 -0500 (CDT) Subject: Re: FreeBSD and Linux SSD Write Speeds From: "Valeri Galtsev" To: "Tim Daneliuk" Cc: "FreeBSD Mailing List" Reply-To: galtsev@kicp.uchicago.edu User-Agent: SquirrelMail/1.4.8-5.el5.centos.7 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 (Normal) Importance: Normal X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 21 Jun 2017 16:28:01 -0000 On Wed, June 21, 2017 11:04 am, Tim Daneliuk wrote: > Disclaimer: Tests below run on lightly loaded systems, but results are ... > surprising: > > Test Case: dd if=/dev/zero of=foo bs=8M count=512 First, I would modify the test and include sync command inside timing, something along these lines: sync; date; dd if=/dev/zero of=foo bs=8M count=512; sync; date ( and your measure is the difference of output of two "date" commands, not what dd reports ) Second, I would make the size dd sends really big, say 10 GB. ( your bs probably _is_ multiple of fs block size, but that is something to watch for too). Next, I would exclude filesystem at all, cut small ( larger than 10 GB) partition off the SSD, and make dd destination that partition, and see what difference does that make. Incidentally, what filesystem is it on FreeBSD side, and how many open files (owned by other processes) do you have on the fiesystems your dd destination file is on? Valeri > > Linux 4.4.0-21-generic on a 2.66GHz Core2 Duo w/8GB memory, older OCZ > SSD/ext4: 310MB/sec writes > > FreeBSD 10-STABLE on an 3.2 GHz Quad Core i5 w/8GB memory, newer Kingston > SSD/ufs: 210MB/sec writes > > Results are repeatable. > > So, what is the likely culprit making FreeBSD 1/3 slower? The FreeBSD > system does does / nfs exported (which I don't quite yet understand since > all the nfs mount points are below it) at the moment, but there is little > or no nfs traffic. > > > _______________________________________________ > freebsd-questions@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-questions > To unsubscribe, send any mail to > "freebsd-questions-unsubscribe@freebsd.org" > ++++++++++++++++++++++++++++++++++++++++ Valeri Galtsev Sr System Administrator Department of Astronomy and Astrophysics Kavli Institute for Cosmological Physics University of Chicago Phone: 773-702-4247 ++++++++++++++++++++++++++++++++++++++++