On Thu, 21 Aug 2003, Madison Kelly wrote:

I think Bill has narrowed the problem to being a fresh-post problem. I replied to his message which came through fine then immidiately sent a repost of a new message I tried to send last night, and it did NOT come through (again).

Without further testing it is hard to say. I still think the problem related to path mtu discovery. Fortunmately this is easy to test. Whoever has access to the mail server just needs to look for the file /proc/sys/net/ipv4/ip_no_pmtu_disc. If the file is not present then pmtu is not running and we don't need to worry about it. If the file does exist, and contains a 0, set it to 1 to turn off pmtu discovery. If the problem goes away it was pmtu discovery afterall. If the problem does not go away then it was not the problem. When I have seen this problem before it has allowed some emails through but not others, even to the same site. Whether an email got through or not was because: (a) it was small enough to not get fragmented at a packet level so pmtu didn't matter or (b) the routing path was changing and so the packets weren't passing through the broken router/firewall. For those interested the problems with pmtu discovery relate largely to misconfigured routers and firewalls, specifically people blocking tcp header info that was "reserved for future use" and then failing to unblock it when the future use came about. Cheers, Rob -- Robert Brockway B.Sc. email: robert-5LEc/6Zm6xCUd8a0hrldnti2O/JbrIOy at public.gmane.org, zzbrock at uqconnect.net Linux counter project ID #16440 (http://counter.li.org) "The earth is but one country and mankind its citizens" -Baha'u'llah -- The Toronto Linux Users Group. Meetings: http://tlug.ss.org TLUG requests: Linux topics, No HTML, wrap text below 80 columns How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml