From owner-freebsd-isp Thu Sep 11 06:14:45 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id GAA20184 for isp-outgoing; Thu, 11 Sep 1997 06:14:45 -0700 (PDT) Received: from burka.carrier.kiev.ua (root@burka.carrier.kiev.ua [193.193.193.107]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id GAA20167 for ; Thu, 11 Sep 1997 06:14:16 -0700 (PDT) Received: (from archer@localhost) by burka.carrier.kiev.ua (8.8.6/8.8.6) id OAA02521 for freebsd-isp@freebsd.org; Thu, 11 Sep 1997 14:14:15 +0300 (EEST) Date: Thu, 11 Sep 1997 14:14:15 +0300 (EEST) From: Alexander Litvin Message-Id: <199709111114.OAA02521@burka.carrier.kiev.ua> To: freebsd-isp@freebsd.org Subject: Re: Quota system is broken? In-Reply-To: X-Newsreader: TIN [UNIX 1.3 unoff BETA 970309; i386 FreeBSD 2.2.1-RELEASE] Sender: owner-freebsd-isp@freebsd.org X-Loop: FreeBSD.org Precedence: bulk > Is the quota system known to be broken? We've tried using it and to say > it's unreliable is an understatement. For example, I have several users > who are completely unrestrained by quota, despite being set up identically > to others where it does work. All the commands, such as repquota, report > they *have* a quota, but they can just keep on writing. On others, I can > set quotas all I want, the system won't recognize them. And in other > accounts, quota reports usages completely wrong, e.g. > Disk quotas for user skunk (uid 1045): > Filesystem blocks quota limit grace files quota limit grace > /usr 127264* 40960 51200 none 325 0 0 > /var 2 1024 10240 2 0 0 > In reality, this user has over 150 megs in /var and only 35 megs in /usr. > Despite being marked as over and no grace left, he can still write in > /usr. > We're running 2.2.2. To the quota topic: I'm not sure that it is quota system bug, but we found here that exeeded quota doesn't prevent user from receiving mail -- mail.local successfully writes to mailbox of such a user. What appeard is that mail.local first opens a mailbox, than changes euid to that of the user, and then writes to mailbox. In such scenario write doesn't fail as it fails when program first changes euid and only than opens a file. Should I patch mail.local, or this is the quota implementation bug? -- Alexander Litvin Duty Postmaster, Lucky Net Ltd.