| |||||||||||||||||||||||||||||||
|
I am using Postfix 2.0.x (yes, I know) on a border mail server, which is configured to relay mail to an internal Exchange server. Normally, incoming mail is tested against the hash defined by relay_domains, as well as the hash defined by relay_recipient_maps, before being sent to the host defined by relayhost. Now that the Postfix primer is done (heh), the problem: one of the domains that my border mail server accepts mail for would like to use their own mail server (i.e. not the one defined by relayhost). My first instinct was simple: set up a transport hash. In my 2.2.x (yes, I know) test environment, I did the following: "/etc/postfix/main.cf" ... transport_maps = hash:/etc/postfix/transport ... "/etc/postfix/transport" other.domain.com relay:[987.65.43.21] # I realise the brackets may not be necessary... I postmap'd the transport file, then restarted postfix (note that the domain already exists in relay_domains, so no changes were made there). I sent a couple of emails through the test environment, and the transport function worked flawlessly. Unfortunately, when I replicated these exact steps on my 2.0.x production servers, the behaviour did not replicate as well. Instead, mail destined for "other.domain.com" continued to be sent to the relayhost, instead of the one defined in the transport file. It would, at first glance, appear as though the transport table is being ignored entirely. I'm not sure what direction to take in terms of further diagnosis, so any insight that the list might be able to provide would be more than welcome. I am happy to post any sort of (sanitised) postconf outputs, and the like, upon request. Thank you. -- _ °v° Daniel Maher /(_)\ Administrateur Système Unix ^ ^ Unix System Administrator "How can a man choose between Fresh and Fly? And believe me, there IS a difference." - Crack Stuntman, 2007.
| ||||||||||||||||||||||||||||||
© 2004-2008 readlist.com