From owner-freebsd-questions@FreeBSD.ORG Mon Jun 23 12:19:31 2003 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C38A937B404 for ; Mon, 23 Jun 2003 12:19:31 -0700 (PDT) Received: from clunix.cl.msu.edu (clunix.cl.msu.edu [35.9.2.10]) by mx1.FreeBSD.org (Postfix) with ESMTP id BF08944003 for ; Mon, 23 Jun 2003 12:19:29 -0700 (PDT) (envelope-from jerrymc@clunix.cl.msu.edu) Received: from clunix.cl.msu.edu (localhost [127.0.0.1]) by clunix.cl.msu.edu (8.12.9/8.12.9) with ESMTP id h5NJJLOg001582; Mon, 23 Jun 2003 15:19:21 -0400 (EDT) Received: (from jerrymc@localhost) by clunix.cl.msu.edu (8.12.9/8.12.9/Submit) id h5NJJFp0001581; Mon, 23 Jun 2003 15:19:15 -0400 (EDT) From: Jerry McAllister Message-Id: <200306231919.h5NJJFp0001581@clunix.cl.msu.edu> To: admin2@enabled.com (admin) Date: Mon, 23 Jun 2003 15:19:15 -0400 (EDT) In-Reply-To: <20030623172734.M94864@enabled.com> from "admin" at Jun 23, 2003 09:27:55 AM X-Mailer: ELM [version 2.5 PL2] MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit cc: freebsd-questions@freebsd.org Subject: Re: newbie: weekly tape backup advice X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 23 Jun 2003 19:19:32 -0000 > > Hi, > > I need some help setting up a tape backup system. I have two FreeBSD machines > and on external SCSI Onstream ADR50. Got any clues how I can start a weekly > back up plan here? It depends a little on the size of your disk compared to your tape capacity. It also depends on how much - amount and frequency - critical data changes. If you can fit everything you want to back up on one tape, just run a full backup (level 0 dump) each time being once per day or once per week or whatever fits your data change pattern. If your critical data change is a lot and a full backup of it will take more than one tape, pick a convenient day of the week and do a full back up and then do incremental backups (level 1 dump) other days. If your disk is so big and the amount of change so much that a week's worth of incremental backup needs more than one tape, then you will want to do a weekly full backup and then increasing levels of incremental back up (level 1 - 6) on the other days. If your amount of data change is quite low - say it is just hosting a fairly static web site and some information database you look at but don't update very often, you might want to consider doing only a weekly full backup or a monthly full backup and weekly incremental backups. Use enough tapes so you are keeping at least three copies of each part of the rotation before reusing a tape. You may also want to do a quarterly or annual archive dump that you store off site and do not reuse for several rotations. For sure, you want to use dump(8). It is part of the system, does the right things with the files and is reliable and doesn't take any tinkering. Unless you have a lot of very critical files open and being changed all the time, don't bother with the warnings about doing a dump on a non-running system. The dump will work just fine. It only means that some file may change between the time the dump started and when it finishes so that file's backup image might not be good. But, if you are doing regular backups -_and not just reusing the same tape all the time_- you will catch that file in a good backup on another day. The man page explains dump pretty well. Mostly you shouldn't need to worry about block size and all the other special stuff. The defaults work best for most circumstances. Determining the media capacity may be the only difficult thing. If one tape will hold the entire backup, just use the '-a' switch. That can work well also for multiple tape dumps with tape drives that give a good end-of-media indication. But some of them - DDS can be an annoying example - tend to not work well when getting near the end of media and will start getting write/read errors before the end-of-media indication actually happens. Then, the system may not handle things very well and you may want to do some calculating and experimenting with either the '-B nnn' parameter or the '-d nnn' and '-s nnn' parameters to specify a media size and force it to change tapes before the problem area is reached. You need to run dump(8) as root. Eventually you will want to not have to retype the dump commands each time or you will want it to run by cron at some time you are not around, so either make a script and run it while su-ed or logged in as root, or make a compiled program that will do the dump calls and make it suid root and then make it owned by root with a group of the only ids that will be allowed to run it and then give it only 750 permissions. One more thing from experience - do not run the head cleaner cartridge any more often than you absolutely have to. In a very clean environment that can actually mean never. But, you will probably need it now and then. Experience will tell when. Those cleaners cause significant wear on the heads and possibly the rest of the mechanism. It doesn't take much to wear those tiny heads down to nothing. So, using them as infrequently as possible will actually help increase head life, not reduce it as some of the accompanying printed material often likes to imply. I think they just way to sell more replacement tape drives. ////jerry > > Thanks in advance, > > Noah >