From owner-freebsd-hackers@FreeBSD.ORG Tue Oct 14 01:38:18 2003 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8DE0916A4B3 for ; Tue, 14 Oct 2003 01:38:18 -0700 (PDT) Received: from c210-49-151-50.thorn1.nsw.optusnet.com.au (c210-49-151-50.thorn1.nsw.optusnet.com.au [210.49.151.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 6900643F93 for ; Tue, 14 Oct 2003 01:38:17 -0700 (PDT) (envelope-from tonymaher@optushome.com.au) Received: from dt.home (localhost [127.0.0.1])ESMTP id h9E8cFDw001142 for ; Tue, 14 Oct 2003 18:38:15 +1000 (EST) (envelope-from tonym@dt.home) Received: (from tonym@localhost) by dt.home (8.12.9p2/8.12.9/Submit) id h9E8cAWe001066 for hackers@freebsd.org; Tue, 14 Oct 2003 18:38:10 +1000 (EST) (envelope-from tonym) Date: Tue, 14 Oct 2003 18:38:10 +1000 (EST) From: Tony Maher Message-Id: <200310140838.h9E8cAWe001066@dt.home> To: hackers@freebsd.org Subject: default route curiosity X-BeenThere: freebsd-hackers@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Technical Discussions relating to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 14 Oct 2003 08:38:18 -0000 Hello, this is just for my own curiosity. On the weekend at work, the comms guys rebuilt a router and our freebsd boxes could not talk to database server in a different subnet for a few hours. The router upgrade failed so upgrade was backed out and routes eventually re-established. All seemed well, could ssh into the freebsd boxes from different subnets but the freebsd boxes could not talk to the database server in one particular subnet (or indeed contact the box on that subnet vi any protocol except ping [icmp]). Other subnet could be reached. sockstat showed there were existing connections to database. The router changed its mac address back and forth during upgrade so did 'arp -ad' but this did not help. Flushing all routes and re-establishing default route and everything immediately worked. The routing table before and after the flushing were identical (and of course default route worked to other subnets before flushing). I assume it was because of the existing database connections there was some bad routing 'info' held in the sytem somewhere. Anyone have any further insight into this. just curious -- tonym