From owner-freebsd-questions Mon Jun 25 7:14:48 2001 Delivered-To: freebsd-questions@freebsd.org Received: from prox.centtech.com (moat2.centtech.com [206.196.95.21]) by hub.freebsd.org (Postfix) with ESMTP id 48CD937B40A for ; Mon, 25 Jun 2001 07:14:44 -0700 (PDT) (envelope-from anderson@centtech.com) Received: (from smap@localhost) by prox.centtech.com (8.9.3+Sun/8.9.3) id JAA12633; Mon, 25 Jun 2001 09:14:37 -0500 (CDT) Received: from sprint.centtech.com(10.177.173.31) by prox via smap (V2.1+anti-relay+anti-spam) id xma012629; Mon, 25 Jun 01 09:14:09 -0500 Received: from centtech.com (proton [10.177.173.77]) by sprint.centtech.com (8.9.3+Sun/8.9.3) with ESMTP id JAA26262; Mon, 25 Jun 2001 09:14:09 -0500 (CDT) Message-ID: <3B374732.E3733672@centtech.com> Date: Mon, 25 Jun 2001 09:14:10 -0500 From: Eric Anderson Reply-To: anderson@centtech.com Organization: Centaur Technology X-Mailer: Mozilla 4.77 [en] (X11; U; Linux 2.2.14-5.0smp i686) X-Accept-Language: en MIME-Version: 1.0 To: Shawn Lussier Cc: freebsd-questions@freebsd.org Subject: Re: NFS & Load Averages References: <20010624235632.R93133-100000@alpha.focalnetworks.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Disk reads are less intensive than NFS writes.. One thing I have found that helps our servers, are these 2 systl switches: vfs.nfs.gatherdelay=0 vfs.nfs.async=1 Seems to boost performance immensely.. Try it on the NFS "server". Eric Shawn Lussier wrote: > > Has anyone any ideas what might be causing one of my systems to have a > high load average than the other system? I am running a recursive cp > between two systems on a 10mbps LAN. The systems, at the time, have > nothing else running aside from the NFS server and client. The first > server, ``omega'', has load averages around .70. This system is a 1.0 ghz > system, with 512 mb ram and 30 gb of ATA/100 disk space. It's running a > 3COM 3c905B card, to my recollection. The second system, ``alpha'' is a > celeron 450 with 256mb ram and 17 gb of ATA/66 storage, running a > ne2000-compatible network card. The load averages on alpha are somewhere > around .12 when doing the copy. The cp process is being run on alpha, > copying the entire /usr tree, which is exported on omega. I am somewhat > perplexed about the cause of the higher load averages on what (should) be > a more powerful system. Again, does anyone have any insights as to the > cause of this? I believe it may be something to do with the disk system on > omega. > > -Shawn L > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-questions" in the body of the message -- ------------------------------------------------------------------------------- Eric Anderson anderson@centtech.com Centaur Technology (512) 418-5792 For every complex problem, there is a solution that is simple, neat, and wrong. ------------------------------------------------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message