Date: Fri, 4 Apr 2003 13:00:38 +1000 From: Christopher Smith <csmith@its.uq.edu.au> To: freebsd-questions@freebsd.org Subject: Re: Weird traceroute problem - SOLVED Message-ID: <9D331AD4-6649-11D7-BABB-000502F96668@its.uq.edu.au> In-Reply-To: <AEBA4362-662A-11D7-BABB-000502F96668@its.uq.edu.au>
next in thread | previous in thread | raw e-mail | index | archive | help
It appears I've been bitten by a bug in the vlan code. I noticed while tcpdumping that the icmp time-exceeded packets were getting back to the vlan parent interface, but not to the vlan interface itself. This thread appears to describe the underlying problem: http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF- 8&threadm=3E05A429.7080506_obluda.cz%40ns.sol.net&rnum=5&prev=/ groups%3Fq%3Dfreebsd%2Bhardware%2Bvlan%2Bbug%26hl%3Den%26lr%3D%26ie%3DUT F-8%26oe%3DUTF- 8%26selm%3D3E05A429.7080506_obluda.cz%2540ns.sol.net%26rnum%3D5 And this PR referenced in it has a patch that fixes the problem: http://www.freebsd.org/cgi/query-pr.cgi?pr=kern/46405 I'm guessing this will only affect some people, as the problem was intermittent (depending on the intervening routers). The ones that were sending back the ICMP packets that were triggering the bug were "Cisco Catalyst 6500s running native IOS" (the networking people here tell me). Presumably these routers change the priority of some ICMP packets ? In any event, can someone please merge the patch in the PR referenced above into the main source tree, because the problem it triggers is rather mystifying :). -- +- Christopher Smith, Systems Administrator ------------------------------+ | Server & Security Group, Information Technology Services | | The University of Queensland, Brisbane, Australia, 4072 | +- Ph +61 7 3365 4046 | email csmith@its.uq.edu.au | Fax +61 7 3365 4065 -+
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9D331AD4-6649-11D7-BABB-000502F96668>