From owner-freebsd-questions Tue Dec 29 16:29:11 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id QAA04796 for freebsd-questions-outgoing; Tue, 29 Dec 1998 16:29:11 -0800 (PST) (envelope-from owner-freebsd-questions@FreeBSD.ORG) Received: from scientia.demon.co.uk (scientia.demon.co.uk [212.228.14.13]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id QAA04575 for ; Tue, 29 Dec 1998 16:29:04 -0800 (PST) (envelope-from ben@scientia.demon.co.uk) Received: from ben by scientia.demon.co.uk with local (Exim 2.10 #3) id 0zv8E0-0004U9-00; Tue, 29 Dec 1998 23:06:52 +0000 Date: Tue, 29 Dec 1998 23:06:52 +0000 From: Ben Smithurst To: "G. Adam Stanislav" Cc: freebsd-questions@FreeBSD.ORG Subject: Re: broken pipe Message-ID: <19981229230652.B17162@scientia.demon.co.uk> References: <3.0.6.32.19981229101429.008a44c0@mail.bfm.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95i In-Reply-To: <3.0.6.32.19981229101429.008a44c0@mail.bfm.org> Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG G. Adam Stanislav wrote: > I was installing via ftp, and sure enough, ran out of disk space. The > installation aborted, and I got a message saying something about a "broken > pipe". I ran the df command and was told my disk was 103% full, and had > negative free space (how can that be???). df reports free space available to normal users. Root however can use more than this, so even if df reports your disk as 100% full, root can still create more files. > Having come from a Windows background, I figured I had to reboot. Sigh. I suppose it's Microsoft's fault, not your fault. > I finally figured that while FreeBSD rarely crashes (unlike Windows > which does it all the time), when it does crash, it seems like the end > of the world. Not really. While my machine has rarely crashed, it comes back up in fairly good condition after. You'll often have to fsck the disks after your machine crashes though, before you can mount them again. > 1. If this happens again (broken pipe), how do I recover? Broken pipe errors occur when a process is writing to a pipe, and they receive a SIGPIPE (I think). If I run something like $ cat /kernel | head -1 > /dev/null I see the error "broken pipe", because head(1) closes its end of the pipe after the first line, but cat(1) tries to keep writing. sh(1) doesn't seem to report "broken pipe" errors like Bash (my login shell) does. > 2. How do I tell the daemons to quit sending all those messages? I can read > them the first time they send them, why do they keep repeating them? Which daemon(s) do you mean, specifically? > 3. Where does a pipe store its temporary data? I'm not sure if a temporary file is created at all, the data may well only be stored in memory. However, the process on the receiving end of the pipe may well want to write data flowing down the pipe onto disk, which may have been your problem. Hopefully someone can clarify this. -- Ben Smithurst ben@scientia.demon.co.uk send a blank message to ben+pgp@scientia.demon.co.uk for PGP key To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message