From owner-freebsd-hackers Mon Jul 10 02:46:20 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id CAA03039 for hackers-outgoing; Mon, 10 Jul 1995 02:46:20 -0700 Received: from gateway.us.sidwell.edu (gateway.us.sidwell.edu [198.3.254.33]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id CAA03032 for ; Mon, 10 Jul 1995 02:46:15 -0700 Received: (from rwatson@localhost) by gateway.us.sidwell.edu (8.6.10/8.6.10) id FAA12694; Mon, 10 Jul 1995 05:46:04 -0400 Date: Mon, 10 Jul 1995 05:46:04 -0400 (EDT) From: Robert Watson X-Sender: rwatson@gateway.us.sidwell.edu To: hackers@freebsd.org Subject: file system full for uid 0 -- doesn't update on file delete Message-ID: Organization: The Sidwell Friends School MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: hackers-owner@freebsd.org Precedence: bulk This may be expected behavior, butI'm not sure, and it cetainly caused me some difficulty ;). I ran ppp in one xterm as a normal user, and dialed to my service provider. I went away for the weekend, left connected. When I came back, I found that the ppp logs in /var/logs had filled my root partition (6 megs of log files for two days?) and I had a long series of log messages saying so. I couldn't figure out at first where my space had gone -- did a du -x, etc, and eventually found the offending file -- I deleted the ppp.log file -- ppp was still running. while du showed that theconsumed disk space was now 12 megs, df showed a 19 meg consumption instead. I kept getting out of disk space errors, and attempting to cat to a file was unpleasent. ;). In the end I rebooted the system. My afterthoughts were.. ;). I shoud have killed ppp, as I assume that the file was being kept open, so the space wasn't being released for some reason (I'm not familiar with the guts of FreeBSD's file handling.) ppp generates too many logs by default ;). that ppp, when run as a normal user, can easily fill up my root partition is a little disturbing. Now presumably if I was really desperate, I could partition cvar onto its own file system, but I'd prefer not to repartition on my home system. It's interesting to wonder whether a long install over a low speed ppp line from the boot floppy would result in an overflowed MFS file system resulting in a halt in the installation or not.. Also, since df and du reported different disk consumption and such, while I'm not suggesting either individual behavior is wrong, consistenc is very nice. Also I wonder what would have happened if ppp had locked in core -- I'd have to reboot to regain disk space.. These are just assumptions based on my assuming that the file being kept open actually is the cause of the problem. Thoughts appreciated.. Robert Watson rwatson@sidwell.edu http://www.sidwell.edu/~rwatson/ The goal of science is to build better mousetraps. The goal of nature is to build better mice.