From owner-freebsd-geom@FreeBSD.ORG Sun Jul 18 22:03:40 2010 Return-Path: Delivered-To: freebsd-geom@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id EB5E71065672 for ; Sun, 18 Jul 2010 22:03:39 +0000 (UTC) (envelope-from freebsd-lists@ideo.com.br) Received: from mail-pw0-f54.google.com (mail-pw0-f54.google.com [209.85.160.54]) by mx1.freebsd.org (Postfix) with ESMTP id CDFCC8FC12 for ; Sun, 18 Jul 2010 22:03:38 +0000 (UTC) Received: by pwj9 with SMTP id 9so1683438pwj.13 for ; Sun, 18 Jul 2010 15:03:38 -0700 (PDT) MIME-Version: 1.0 Received: by 10.142.164.3 with SMTP id m3mr5502383wfe.297.1279489277054; Sun, 18 Jul 2010 14:41:17 -0700 (PDT) Received: by 10.142.170.6 with HTTP; Sun, 18 Jul 2010 14:41:17 -0700 (PDT) Date: Sun, 18 Jul 2010 18:41:17 -0300 Message-ID: From: Marco Haddad To: freebsd-geom@freebsd.org Content-Type: text/plain; charset=ISO-8859-1 X-Content-Filtered-By: Mailman/MimeDel 2.1.5 Subject: gvinum performance issues X-BeenThere: freebsd-geom@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: GEOM-specific discussions and implementations List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 18 Jul 2010 22:03:40 -0000 Hello, I've seen somewhere a thread about this topic, but I can't find it here, where I think it ought to be. So, let me describe what is going on: I had this Core2 Quad running as Internet gateway, web server and file server, doing fine with FreeBSD 7.2 for a long time. The problem came with the upgrade to version 8.0. It is composed of two gmirror volumes, each build up from a pair of 500GB disks. They are not the problem. Because of the nature of this server role, a long history of backups is necessary, and for that task there was a single 1.5TB disk, which got filled up after some time. Then I decided to get two more 1.5TB disks and use gvinum to build a raid5 volume for the backup. That was done with success. The backup operation that used to take four hours to complete, now was taking 6 hours, what was ok for an over night backup. Then I decided to upgrade to version 8.0, so I could grow the volume just adding more disks, without losing its content. The upgrade went well with no problems at all, but the backup was already running for three days when I realized that something was wrong. The transfer rate was around 200KB/s, a fraction of what it was with version 7.2. After trying all possible solutions found, including rebuilding parity, destroying the volume and recreating under version 8.0, I decided the just concat the three disks and let the backup roll, while I seek for help. I read somewhere Ulf telling that he had received a number of similar complaints, but there was no solution on that thread yet. So if someone here had this problem solved, please enlighten me. Thanks in advance, Marco Haddad