From owner-freebsd-questions@FreeBSD.ORG Wed Oct 22 03:03:17 2014 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 5C480F1B for ; Wed, 22 Oct 2014 03:03:17 +0000 (UTC) Received: from server1.shellworld.net (shellworld.net [69.60.117.94]) by mx1.freebsd.org (Postfix) with ESMTP id 38F6D791 for ; Wed, 22 Oct 2014 03:03:16 +0000 (UTC) Received: from server1.shellworld.net (localhost [127.0.0.1]) by server1.shellworld.net (Postfix) with ESMTP id 10BBB22956 for ; Tue, 21 Oct 2014 23:03:10 -0400 (EDT) To: freebsd-questions@freebsd.org Subject: Network Interface Keeps Dying FreeBSD9.1 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <95467.1413946989.1@server1.shellworld.net> Date: Tue, 21 Oct 2014 22:03:10 -0500 From: "Martin G. McCormick" Message-Id: <20141022030310.10BBB22956@server1.shellworld.net> X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.18-1 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 22 Oct 2014 03:03:17 -0000 This system has 4 NIC's. Only one of the NIC's is likely to ever be used. When setting up the system, we selected DHCP for now and accidentally discovered that we had picked the wrong Ethernet jack so the interface configuration attempt failed when we tried it. We went ahead and finished configuring the system and I used ifconfig to basically buzz out which interface came to life when an Ethernet cable was plugged in. They all show up with gbe2, gbe3, gbe0 and gbe1 as one plugs in to jacks 1, 2, 3 and 4. We wanted gbe0 so I started the system with the other 3 unconnected. You've never seen so much flapping of the connected interface. Line after line of gbe0 being down followed by gbe0 being up. Eventually, gbe0 would stop flapping and, no surprise, was either showing "no carrier" or 0.0.0.0 as the IP address. I got a bigger hammer, so to speak and manually forced it to the IP address we needed it to be along the line of defaultrouter="192.168.255.254" ifconfig_gbe0="inet 192.168.255.20 netmask 255.255.255.0" For the rest of the interfaces, I have ifconfig_gbe1="down" and the same for 2 and 3. Starting up, we have more flapping but when things settle down, one can actually ping various other addresses in that and other subnets in our network so it's finding the router and we are getting ICMP returns. As soon as I actually try to do anything such as scp something from another system, gbe0 collapses with /var/log/messages reporting only that gbe0 went down. The watchdog timer springs in to action and gbe0 is back again for all the pinging one could ever want to do. Has anybody seen this behavior? Sysinstall has a function called netInteractive that will re-run the network setup but I haven't been able to make it happen and re-running sysinstall without arguments doesn't seem to let one re-do the network setup. This is a new system that has never worked yet but there is no randomness to the interface crashing. It crashes every time one either contacts it from another system or it contacts another system for anything other than ICMP. Any constructive suggestions are welcome. Martin McCormick