Prioritizing among MX records

[phn at icke-reklam.ipsec.nu]

Barry Margolin <barmar at alum.mit.edu> wrote:
> In article <c2i7rv$le4$1 at sf1.isc.org>, phn at icke-reklam.ipsec.nu wrote:

>> Under normal circumstances *some* mail will be delivered=20
>> to "higer-MX" due to various sources ( network disturbances etc),
>> the amount seems to depend on the "diversity" of the mailhosts.

> This should be very rare -- a network disturbance has to be pretty 
> severe for a connection attempt to fail completely.

It's surpricingly common. slightly less then a percent if my 
memory serves me right.

Part of explanation might be that the hosts experienced 
losses for "distant nets", another that their nameservers
already cached the us mailserver as they might have received
mail from that ( sending mail was done by geografically 
closest mailserver, the us senders would prefer the us 
mailserver(but fall back by mx used on the inside).
Other explanations are welcome, i was puzzled myself but 
as it was consistent and quite stable i regarded this 
as "one of the unsolved mysteries of life (Tm)"

>> I have observed in one organisation that has MX both in Europe
>> and US that US senders seems to prefer the US mailhost in spite
>> of "higer-mx".

> Are you sure they don't have geographically disperse DNS servers that 
> give out different MX records, i.e. the DNS server in US has lower MX 
> preferences on the US servers?

On that part i'm certain that DNS was consistent. I did the dns myself !

> -- 
> Barry Margolin, barmar at alum.mit.edu
> Arlington, MA
> *** PLEASE post questions in newsgroups, not directly to me ***


-- 
Peter Håkanson         
        IPSec  Sverige      ( At Gothenburg Riverside )
           Sorry about my e-mail address, but i'm trying to keep spam out,
	   remove "icke-reklam" if you feel for mailing me. Thanx.