On Wed, 3 Sep 2003, Dmitry Sergienko wrote:
> relay2 is machine where delivery is being done, so I don't understand
> how it can be in retry database or inaccessible host. Maybe you meant
> relay1? If so, it is in retry database, because it is inaccessible.
With respect, if a host is inaccessible, it shouldn't be mentioned in an
MX list because that causes any client that is sending to the domain to
do unnecessary work.
> But the main problem is that I don't want exim to try any other MX'es in
> this case. Otherwise I have a ping-pong beetween relay2 and relay3.
>
> That's a log of such delivery:
The log shows no mention of relay2 at all. It just shows relay1, relay3,
and relay 4:
> 78372 routing test@???
<snip>
> 78372 --------> dnslookup router <--------
<snip>
> 78372 host_find_bydns yield = HOST_FOUND (2); returned hosts:
> 78372 mail.relay1.example.net xxx.xxx.xxx.242 5
> 78372 relay3.example.net xxx.xxx.xxx.65 25
> 78372 relay4.example.net xxx.xxx.xxx.119 150
Then Exim tries to deliver to those hosts, exactly according to the MX
values:
> 78372 >>>>>>>>>>>>>>>> Remote deliveries >>>>>>>>>>>>>>>>
> 78372 --------> test@??? <--------
> 78373 mail.relay1.example.net [xxx.xxx.xxx.242] status = unusable
... because it has no doubt tried it before.
> 78373 Connecting to relay3.example.net [xxx.xxx.xxx.65]:25 ... connected
And then it delivers the message to relay3.
I can't see any problem in Exim's behaviour here...
Philip
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
Get the Exim 4 book: http://www.uit.co.uk/exim-book
