From owner-freebsd-ports@FreeBSD.ORG Thu Jul 28 23:54:54 2005 Return-Path: X-Original-To: freebsd-ports@freebsd.org Delivered-To: freebsd-ports@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 616F816A41F; Thu, 28 Jul 2005 23:54:54 +0000 (GMT) (envelope-from lavalamp@spiritual-machines.org) Received: from mail.digitalfreaks.org (arbitor.digitalfreaks.org [216.151.95.158]) by mx1.FreeBSD.org (Postfix) with ESMTP id 1473F43D49; Thu, 28 Jul 2005 23:54:53 +0000 (GMT) (envelope-from lavalamp@spiritual-machines.org) Received: by mail.digitalfreaks.org (Postfix, from userid 1022) id 0C45131980B; Thu, 28 Jul 2005 19:54:53 -0400 (EDT) Received: from localhost (localhost [127.0.0.1]) by mail.digitalfreaks.org (Postfix) with ESMTP id 0AEE131F406; Thu, 28 Jul 2005 19:54:53 -0400 (EDT) Date: Thu, 28 Jul 2005 19:54:52 -0400 (EDT) From: "Brian A. Seklecki" X-X-Sender: lavalamp@arbitor.digitalfreaks.org To: freebsd-ports@freebsd.org Message-ID: <20050728193334.P7262@arbitor.digitalfreaks.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Cc: Phil Homewood , jeh@FreeBSD.org Subject: misc/amanda / Users X-BeenThere: freebsd-ports@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting software to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 28 Jul 2005 23:54:54 -0000 all, re: ports/73956 ...although the approach used in this PR is a great improvement, should we really be defaulting to using 'operator:backup' if no user is set? Won't that encourage people to unlock the operator account by assigning it a valid shell, or even a password? It's unlikely people will take the high ground and execute all Amanda commands from sudo(8). Aren't there hooks for creating psuedo accounts in Ports? For example, in NetBSD pkgsrc there's a PKG_USERS and PKG_GROUPS that can be assigned with low UID values. If so, why not default to creating an 'amanda' or 'backup' user in the secondary group operator? It's just that Amanda has some serious fudemental security issues as it is (no offense to them, it works well), such RHosts style authentication, depedency on inetd/xinetd, and lack of inline network encryption. I just think we should be more proactive; I think even recent versin of Redhat ship it with an amanda user. --- Also, we should probably add a pkg-message for the client and server mentioning required entries in inetd.conf(5), or is the thinking here that Amanda is so involved that people are going to refer to the docs anyway? P.S., this would be an excellent use for the IPSec hooks in inetd(8). ~BAS l8* -lava