Szerző: Giuliano Gavazzi Dátum: Címzett: exim-users Tárgy: Re: [Exim] temporary error taken as permanent error?
Ok, I searched my list archives and think I have the reason of the
immediate failure on a temp error:
the MX in question, mx5.iol.it, was contacted and failed over 4 days
before this temporary failure and there had been no intermediate
successes (or even attempts):
grep mx5.iol.it exim_mainlog
...
2003-06-05 00:59:35 HFZECO-000HCH-R8 SMTP error from remote mailer
after initial connection: host mx5.iol.it [193.70.192.92]: 421
smtp8.libero.it Service not available - too busy
2003-06-05 00:59:35 HFZECO-000HCH-R8 == @iol.it R=dnslookup
T=remote_smtp defer (0): SMTP error from remote mailer after initial
connection: host mx5.iol.it [193.70.192.92]: 421 smtp8.libero.it
Service not available - too busy
...
2003-06-11 12:11:36 HGBDR7-0004RQ-KG SMTP error from remote mailer
after initial connection: host mx5.iol.it [193.70.192.90]: 421
smtp7.libero.it Service not available - too busy
[and then, after attempts to the other IPs of mx5, the failure]
I can understand why this happened and the reason of this per host
behaviour, but since the domain has more that one MX, and for at
least one of these MXes there had been one successful delivery on
2003-06-10, why not try that MX, or any other one, and in case of
failure, requeue the message since the cutoff time had not been
reached yet for one of these?
In other words, why stop at one MX when there are 5?
[I ignore here that for this domain there are five MXes, mx1 to mx5,
with 5 addresses each from the same pool of 5 addresses. In other
words each mx1 to mx5 correspond to the same five machines...
probably a method to spread the load whatever method the clients use.]