From owner-freebsd-stable@FreeBSD.ORG Sat Aug 9 18:37:24 2008 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id A53A5106568C for ; Sat, 9 Aug 2008 18:37:24 +0000 (UTC) (envelope-from cliftonr@lava.net) Received: from outgoing01.lava.net (cake.lava.net [IPv6:2001:1888:0:1:230:48ff:fe5b:3b50]) by mx1.freebsd.org (Postfix) with ESMTP id 200FF8FC15 for ; Sat, 9 Aug 2008 18:37:24 +0000 (UTC) (envelope-from cliftonr@lava.net) Received: from malasada.lava.net (malasada.lava.net [64.65.64.17]) by outgoing01.lava.net (Postfix) with ESMTP id 1FF7FD0072; Sat, 9 Aug 2008 08:37:23 -1000 (HST) Received: by malasada.lava.net (Postfix, from userid 102) id 66661153882; Sat, 9 Aug 2008 08:37:22 -1000 (HST) Date: Sat, 9 Aug 2008 08:37:22 -1000 From: Clifton Royston To: KES Message-ID: <20080809183721.GA9982@lava.net> Mail-Followup-To: KES , freebsd-stable@freebsd.org References: <358831218288212@webmail24.yandex.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <358831218288212@webmail24.yandex.ru> User-Agent: Mutt/1.4.2.2i Cc: freebsd-stable@freebsd.org Subject: Re: Fwd: IMPORTANT! Network is unreachable X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 09 Aug 2008 18:37:24 -0000 On Sat, Aug 09, 2008 at 05:23:32PM +0400, KES wrote: > 09.08.08, 16:22, "Matthew Seaman" : > > Andrew Snow wrote: > > > Usually if there is more than IP in a given subnet on an interface, you > > > give it a /32 netmask. Only the first IP in a subnet should have the > > > full netmask. > > > > > > So your example should look like this: > > > > > > inet 10.11.16.14 netmask 0xffffff00 broadcast 10.11.16.255 > > > inet 10.11.16.9 netmask 0xffffffff broadcast 10.11.16.9 > > /32 netmasks for 2nd and subsequent IP alias addresses used to be > > mandatory and are arguably more correct, but nowadays you can use > > the actual netmask for the network instead. Was fixed a year or > > two ago. It's a wetware compatibility thing -- other unixoid OSes > > never had the /32 netmask requirement, and it kept tripping people up > > when swapping between OSes. > > Unfortunately I can't say exactly what the problem the OP is experiencing > > is due to, but the way routes are appearing and disappearing on a 5 > > minute timescale does suggest dynamic routing problems to me. As a > > work-around, if the OP wanted to override the information routed gets > > from the network, then he could use /etc/gateways to have the local > > routed append some static routes to the routing table -- see routed(8) > > for the gory details. Losing a route for a directly attached network > > looks like a bug to me though. ... > > > > inet 10.11.16.14 netmask 0xffffff00 broadcast 10.11.16.255 > > > inet 10.11.16.9 netmask 0xffffffff broadcast 10.11.16.9 > /24 mask on each IPs on same interfaces is working fine on FreeBSD 6.3 > So I do not think that problem is with the network mask. Because of even ping 10.11.16.14 > returns network is unreachable! > Now when I upgraded to v7 I see trouble described earlier. > So this is must be counted as BUG of v7 I happened to see recently a report of a similar problem with 7.0 on a private mailing list. Again, there were multiple IP addresses configured within the main subnet of the interface (this time configured as /32s on other physical interfaces) and again, after a while the system lost connectivity to its main subnet and "forgot" how to ARP for addresses on the interface. An important similarity - the routing info like yours showed the attached network with the G flag, as being reachable via the gateway address within the same subnet. I can't troubleshoot this, no access to the system in question, but I thought it might help to know that others have run into the same problem. > The thing which is very interesting is: > Why period is 5 min? Might be something to do with ARP? Not sure. -- Clifton -- Clifton Royston -- cliftonr@iandicomputing.com / cliftonr@lava.net President - I and I Computing * http://www.iandicomputing.com/ Custom programming, network design, systems and network consulting services