From owner-freebsd-net@FreeBSD.ORG Mon Feb 18 15:51:44 2008 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 96B9616A417 for ; Mon, 18 Feb 2008 15:51:44 +0000 (UTC) (envelope-from lup@apriko.ru) Received: from mail.apriko.ru (ns1.apriko.ru [91.198.212.4]) by mx1.freebsd.org (Postfix) with ESMTP id 4AAF613C458 for ; Mon, 18 Feb 2008 15:51:44 +0000 (UTC) (envelope-from lup@apriko.ru) Received: from localhost (heart [127.0.0.1]) by mail.apriko.ru (Postfix) with ESMTP id CDBDCC896AA for ; Mon, 18 Feb 2008 18:19:12 +0300 (MSK) X-Virus-Scanned: amavisd-new at apriko.ru Received: from mail.apriko.ru ([127.0.0.1]) by localhost (mail.apriko.ru [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id rqncAo1IjHTN for ; Mon, 18 Feb 2008 18:19:07 +0300 (MSK) Received: from [127.0.0.1] (unknown [91.198.212.73]) by mail.apriko.ru (Postfix) with ESMTP id 25FBBC888EA for ; Mon, 18 Feb 2008 18:19:07 +0300 (MSK) Message-ID: <47B9A1DD.5040408@apriko.ru> Date: Mon, 18 Feb 2008 18:18:53 +0300 From: Alexander Pavlov User-Agent: Thunderbird 2.0.0.9 (Windows/20071031) MIME-Version: 1.0 To: freebsd-net@freebsd.org Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Subject: mpd-netgraph related issue X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 18 Feb 2008 15:51:44 -0000 Hello, I'm using a couple of freebsd boxes for providing pptp vpn services. Some of them have nearly identical hardware and software configuration. Others don't (I use amd64 and i386 architectures). But the same problem is observed on either of them. The problem is that at arbitrary moment of time (almost daily) box losses connectivity. I've tried to upgrade from 6.1-RELEASE to 6.2-RELEASE, but problem remains. Now I'm running: #uname -v FreeBSD 6.2-RELEASE-p9 #1: Sat Jan 26 17:31:44 UTC 2008 lup@torment.skynet:/usr/obj/usr/src/sys/CLONED-CORE # mpd4 -v Version 4.3 For NAT I use ipnat: #ipf -V ipf: IP Filter: v4.1.13 (416) Two em-NIC's are on board # pciconf -l|grep em em0@pci4:0:0: class=0x020000 card=0x348d8086 chip=0x108c8086 rev=0x03 hdr=0x00 em1@pci5:5:0: class=0x020000 card=0x348d8086 chip=0x10768086 rev=0x05 hdr=0x00 I observe three different scenarios: 1. Complete loss of connectivity (box is working). Box itself is not frozen, but I can't even ping localhost. I'm not running outof mbufs -- output of netstat -m seems to be ok. Other kernel limitations are not exceeded also (but not sure that I've checked all of them). 2. Complete loss of connectivity (box hangs up). Symtoms are same, except a very strange behaviour of box. When I'm trying to log in locally, I do not observe any reaction on my keystrokes, but when I'm trying to shutdown box, it says acpi: suspend request ignored (not ready yet) and shows me everything I typed. Using power button as enter key %)) I can login locally and run some commands. top shows load averages beyond 1, but no network connectivity though. 3. Particial loss of connectivity. I can connect to ports 22, 1723 and other opened ports via LAN interface (on which mpd operates) using telnet. ICMP also works. But I can't login using ssh and I can't connect using pptp client. ssh via wan interface works perfectly. There are some symptoms, which are same for all of the scenarios above: 1. I can't shutdown mpd. Sending KILL signal does nothing. 2. When I try to run ngctl (e.g. ngctl types), it hangs, but I can kill ngctl process. 3. When I try to reboot box, it hangs (but still responds to icmp echo requests) until I hard-reset it. Can anybody explain this abnormal behaviour?