From owner-freebsd-current Thu Oct 17 01:32:10 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id BAA15815 for current-outgoing; Thu, 17 Oct 1996 01:32:10 -0700 (PDT) Received: from quagmire.ki.net (root@quagmire.ki.net [205.150.102.1]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id BAA15776 for ; Thu, 17 Oct 1996 01:32:06 -0700 (PDT) Received: from spirit.ki.net (root@spirit.ki.net [205.150.102.51]) by quagmire.ki.net (8.7.5/8.7.5) with ESMTP id EAA06781 for ; Thu, 17 Oct 1996 04:32:15 -0400 (EDT) Received: from localhost (scrappy@localhost) by spirit.ki.net (8.7.5/8.7.3) with SMTP id EAA05418 for ; Thu, 17 Oct 1996 04:32:08 -0400 (EDT) X-Authentication-Warning: spirit.ki.net: scrappy owned process doing -bs Date: Thu, 17 Oct 1996 04:32:02 -0400 (EDT) From: "Marc G. Fournier" To: current@freebsd.org Subject: Iozone: local vs nfs drives Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-current@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Hi... I was trying to prove a point in news tonight between using an nfs mounted drive vs a local drive for a news server (one of the gentleman doesn't seem to believe there is a difference). My argument involved using iozone to show the throughput differences between nfs/local r/w's, and the results I got don't look right: this is my nfs mounted drive (133k/s and 382k/s): IOZONE performance measurements: 133152 bytes/second for writing the file 382386 bytes/second for reading the file vs local drive (1.78M/s and 938k/s): IOZONE performance measurements: 1784080 bytes/second for writing the file 938585 bytes/second for reading the file They look right in so far that I would expect the local file system to be faster then the remote...but I would also have expected that reading the local file would be faster the writing it, as was reflected in the nfs-mounted file system... Both systems are running 2.2-Current, same kernels, about 3 days old... I've tried varying the size of file iozone uses, up to a 30Meg file, and the results seem to be quite consistent: IOZONE performance measurements: 1984490 bytes/second for writing the file 1117239 bytes/second for reading the file Why would reading the file be half as fast as writing it? Marc G. Fournier scrappy@ki.net Systems Administrator @ ki.net scrappy@freebsd.org