From owner-freebsd-questions@FreeBSD.ORG Mon Dec 12 18:42:11 2005 Return-Path: X-Original-To: freebsd-questions@freebsd.org 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 CEA0216A41F for ; Mon, 12 Dec 2005 18:42:11 +0000 (GMT) (envelope-from kdk@daleco.biz) Received: from ezekiel.daleco.biz (southernuniform.com [66.76.92.18]) by mx1.FreeBSD.org (Postfix) with ESMTP id 37F7443D5A for ; Mon, 12 Dec 2005 18:42:10 +0000 (GMT) (envelope-from kdk@daleco.biz) Received: from [192.168.2.2] ([69.27.149.254]) by ezekiel.daleco.biz (8.13.1/8.13.1) with ESMTP id jBCIeqIU037652; Mon, 12 Dec 2005 12:41:15 -0600 (CST) (envelope-from kdk@daleco.biz) Message-ID: <439DC428.9060105@daleco.biz> Date: Mon, 12 Dec 2005 12:40:40 -0600 From: Kevin Kinsey User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.7.12) Gecko/20051026 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Sasa Stupar References: <000601c5ff28$9fb71e00$46933c50@Medion> <20051212144948.GB2325@flame.pc> <17309.37575.337162.532084@jerusalem.litteratus.org> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-questions@freebsd.org Subject: Re: Slices X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 12 Dec 2005 18:42:12 -0000 Sasa Stupar wrote: > You can also do the same as I did. I have just configured one slice > named / which > takes all the space on hdd and now I don't need to worry about space > shortage. OK then, but suppose we have some runaway process logging errors to syslogd at some absurb rate (n per second). Suppose you aren't sitting at the console or some terminal running "df -h" at the moment. Maybe it's that somebody came into the office on Saturday and the line printer was off-line, but they didn't know it and sent it a job. Or, perhaps someone backs up a big bunch of stuff to some NFS or SMBFS mount point that happens to not be mounted at the time because some secretary turned off her box or a janitor ran a mop over some plug somewhere. Maybe your 'Net connection, on one side or the other, takes a nose dive, but things appear OK locally, and John in marketing finishes up the latest sales e-mail and sends it to all 250,000 potential clients. Perhaps a junior sysadmin decides to use his homedir as a temporary FTP repository for his mp3 collection, while he's headed to Mom and Dad's for the holidays, and forgets that he's got a tree of movies in that dir as well. There are several possibilities, and sooner or later, somewhere, someone will get caught be by one of them. Suppose then that syslogd fills up /var/log/ by writting a gazillion MB to messages ( ... messages0, messages1 ... messages n) and, for whatever reason, say, it's not time, newsyslog hasn't rotated the logfiles yet. Or Sendmail panics because it can't write to the spool, 'cause it's full (and then starts writing error messages to syslogd....) What I'm saying is: before you can know it, once in a while, your / (in the setup you describe) can end up with no space left. How will you recover now? Your hard disk is full; if you are fortunate, you'll be able to get up in single-user from the console. Of course, you could be a thousand miles away.... That's why my /var is seperate from my / .... There may be someone who can shoot my scenario relatively full of holes, but there *was* a reason why we have a recommended paritioning scheme with a seperate /, /var, /usr, and so on, and I'm not at all sure that it's time to abandon it entirely. My $0.02, Kevin Kinsey -- I need another lawyer like I need another hole in my head. -- Fratianno