Warner Losh writes:
> I have been having MX problems that I can't seem to track down.
> What happens is that I have the following MX setup:
> harmony 5
> rover 10
> mailhub 20
> rover is running exim. When harmony goes down, mail makes it as far
> as rover and then loops infintely (well to 30) and bounces. When
> harmony is up, it gets the mail. My question is how do I find out why
> exim is repeatedly sending mail to itself? I asked once before, but
> didn't get any reply.
>
> Many thanks for any help or assistance you might be able to render me
> here. I'm using 1.60, if that matters from the FreeBSD ports
> collection, but would be happy to upgrade to a newer "stable" version
> if that would fix my problem.
A host should never send a message to a host with MX values >= its own.
There *were* bugs in Exim in this area, but only in the case of multiple
MX records of the same priority, and I think they were fixed by 1.60 anyway.
It seems possible that something is wrong with rover's self-recognition,
so that it doesn't realise that the second MX record refers to it.
Try submitting a message of the the offending sort with
exim-pathname -d9 [usual sendmail options, e.g. -t] < sample-message
which should continue to trace activities through the first delivery attempt.
Alternatively, get it queued without a transmission attempt by using -odq,
freeze it with "exim -Mt msgid", and then at your leisure debug a
transmission attempt with
exim-pathname -d9 -M msgid
This may just show you exim making the loopback call to rover without any
very good explanation as to why. What you really need to know is what
host_find_interfaces() is returning, and why it doesn't contain the address
that it is sending these messages to itself on, supposing that that is
indeed the source of the problem.
Chris Thompson Cambridge University Computing Service,
Email: cet1@??? New Museums Site, Cambridge CB2 3QG,
Phone: +44 1223 334715 United Kingdom.
