On Wednesday 29 October 2003 12:26, Mike Mueller wrote: > On Wednesday 29 October 2003 12:05, Jason Tower wrote: > > the catch was, > > they wanted to have zero downtime when they made the DNS changes so > > the web server had to function properly on both T1s simultaneously > > (the two T1s were from different ISPs and had totally different > > subnets). this is harder than it seems at first glance; in order > > for this to function you have to make sure that traffic "leaves" on > > the same interface it arrived on, regardless of which route is > > preferred. > > I stumbled on the zero downtime and iso-interface enter/exit > requirements; if one of the T1 interfaces goes down, what happens to > responses to queries that entered on the downed interface? are they > abandoned or rerouted to the other interface?
even if they were re-routed thru the still active interface, it would probably break the tcp connection since the host at the other end would see a different source IP. however, this is only conjecture since i have not tried it. jason -- TriLUG mailing list : http://www.trilug.org/mailman/listinfo/trilug TriLUG Organizational FAQ : http://trilug.org/faq/ TriLUG Member Services FAQ : http://members.trilug.org/services_faq/ TriLUG PGP Keyring : http://trilug.org/~chrish/trilug.asc
