From owner-freebsd-questions@FreeBSD.ORG Fri Nov 28 07:42:48 2003 Return-Path: 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 B3D9216A4CE for ; Fri, 28 Nov 2003 07:42:48 -0800 (PST) Received: from cactus.fi.uba.ar (cactus.fi.uba.ar [157.92.49.108]) by mx1.FreeBSD.org (Postfix) with ESMTP id A253B43F75 for ; Fri, 28 Nov 2003 07:42:46 -0800 (PST) (envelope-from fgleiser@cactus.fi.uba.ar) Received: from cactus.fi.uba.ar (cactus.fi.uba.ar [157.92.49.108]) by cactus.fi.uba.ar (8.12.3/8.12.3) with ESMTP id hASFgStf071001 for ; Fri, 28 Nov 2003 12:42:28 -0300 (ART) (envelope-from fgleiser@cactus.fi.uba.ar) Date: Fri, 28 Nov 2003 12:42:28 -0300 (ART) From: Fernando Gleiser To: questions@freebsd.org Message-ID: <20031128122236.F70945-100000@cactus.fi.uba.ar> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Scanned-By: MIMEDefang 2.37 Subject: Problems with mpd in pptp mode X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 28 Nov 2003 15:42:48 -0000 X-List-Received-Date: Fri, 28 Nov 2003 15:42:48 -0000 X-List-Received-Date: Fri, 28 Nov 2003 15:42:48 -0000 I'm trying to set up mpd as a pptp client to a remote server. It authenticates right and setts up the tunnel, but it doesn't quite work. I narrowed the problem to sone nasty routing problem, because the pptp server's IP, as defined in mpd.links is the same as the remote ip of the ng1 interface. In detail: my client: dynamic IP given my my ISP. pptp server: a.b.c.d once I run mpd, I get the following: ng1: flags=88d1 mtu 1458 inet a.b.e.f --> a.b.c.d netmask 0xffffffff inet6 fe80::250:daff:fe76:62e5%ng1 prefixlen 64 scopeid 0x9 as you see, the pptp server's ip is the same as the remote ptp one. After that, the box tries to route all traffic to a.b.c.d via ng1 including the gre packets and the ones before encapsulation. here's a tcpdump snippet of what happens: bash-2.05b# tcpdump -nli ng1 tcpdump: listening on ng1 12:22:59.027790 external > a.b.c.d: gre [KAv1] ID:d03c A:15 [|gre] 12:23:02.587838 external > a.b.c.d: gre [KAv1] ID:d03c A:16 [|gre] 12:23:02.820064 a.b.e.f > a.b.c.d: icmp: echo request 12:23:03.827900 a.b.e.f > a.b.c.d: icmp: echo request 12:23:04.837908 a.b.e.f > a.b.c.d: icmp: echo request 12:23:05.137864 external > a.b.c.d: gre [KAv1] ID:d03c A:17 [|gre] 12:23:05.847928 a.b.e.f > a.b.c.d: icmp: echo request 12:23:06.857961 a.b.e.f > a.b.c.d: icmp: echo request 12:23:07.867965 a.b.e.f > a.b.c.d: icmp: echo request 12:23:08.027909 external > a.b.c.d: gre [KAv1] ID:d03c A:18 [|gre] 12:23:08.877974 a.b.e.f > a.b.c.d: icmp: echo request 12:23:09.887987 a.b.e.f > a.b.c.d: icmp: echo request The icmp are fine, but the gre packets shouldn't be there. It seems to me I should delete the actual route to a.b.c.d and replace it with a route via my actual default gateway. That way, all packets to the remote peer would go out via xl0 (external interface) and no via ng1 (the tunnel), but I'm not sure about how I can do it. Thanks in advance. Fer