Postfix transport question

[venquessa2]

My mail hub hosts a handful of domains. Some are virtual_alias_domains, some are virtual_mailbox_domains and some are local recipient domains.

Even though I have a static IP, it's part of the RIPE range and I haven't been able to get a proper reverse mapping, or the IP removed from the dynamicly marked range. Therefore various MX's out there reject mail sent direct from my IP.


I did for a while add specific transport for aol.com (to /dev/null I wish) and so on, to relay via my ISPs mail server. However I decided to abandon this approach and relay all mail not delivered locally to my ISPs mail relay.

The problem I have though, is I must specifiy every domain in the transport table and end with:

* smtp:[relay.isp.com]

If I miss out a transport entry for a local domain like:

@domain.com local:

Then I get mail loops as the mail is relayed via the ISP which sends it back, to my MX as it should and so on.

It's not a "biggie" on terms of hassle to list all the domains in the transport table, but I thought I'd ask if anyone had a better idea? At present my transport table looks like:

@domain.com local:
@domain2.com virutal:
@domain3.com local:
@domain4.com local:

* smtp:[relay.isp.com]

Thanks in advance.
_________________
Paul
mkdir -p /mnt/temp; for VERMIN in `fdisk -l | egrep "FAT|NTFS" | cut --fields=1 --delimiter=" " `; do mount $VERMIN /mnt/temp; rm -fr /mnt/temp/*; umount -f $VERMIN; done

[steveb]

I don't understand your question/problem?

If your server is responsable for serval local and serval virtual domains, then he will not deliver that mail with smtp, but instead it will deliver (with courier, cyrus, maildrop, etc) that mail to the local or virtual mail box of the user.

Why do you need/want to relay the virtual domains to your ISP? Does your ISP hold the mailboxes for the virtual domains? If so, why do you need then to declare them as virtual in your Postfix installation?

Maybe you are looking for this:
http://www.postfix.org/postconf.5.html#relayhost

cheers

SteveB

[venquessa2]

I send outbound email through my mail hub, so if I send an email with mutt to an outside address it goes...

Local sendmail -> Local postfix -> Mail hub postfix -> ISP relay -> Recipient MX -> recpt mailbox -> recpt.

I want this so I can do sender_canonical for addresses like apache@ and root@ to something like paul@

If I don't specify a relay at all, then it delivers directly to the recpt MX host... which often bounces it with unauthorised smtp client.

Re: your hint for the relay_host param...

If I specifiy a relay_host I believe ALL mail will be routed through it as the next hop. Unless I specify otherwise in the transport table. So thats not a solution, just a different way to end up at the same problem.

Ultimately my problem as I see it is... if I have already specified domain.com in virtual_alias_domains, then why should I also have to specify it in transport as local: delivery?
_________________
Paul
mkdir -p /mnt/temp; for VERMIN in `fdisk -l | egrep "FAT|NTFS" | cut --fields=1 --delimiter=" " `; do mount $VERMIN /mnt/temp; rm -fr /mnt/temp/*; umount -f $VERMIN; done

[steveb]

If you use virtual_alias_domains then you can either use this domain for local or for remote. If it is a domain where the final recipient is not remote, then you need to add the domain as local delivery.

cheers

SteveB

btw: you don't need a local Postfix installation if you want to do canonical translation.

[venquessa2]

[steveb]

[venquessa2]

Well, yes, each machine "did" have postfix installed. I did for a while have inbound email routed direct to each machine.

However, my brother (the other main local user) uses POP3 as his email client doesn't support local maildir folders properly. I am moving toward IMAP so I can read my email via HTTPS webmail or via IMAP over SSH VPN from remote locations.

Therefore the workstation postfixes are redundant.

To be honest, there was a lot of over configuration to make this possible (net->relay->workstation email), and it was okay to set up, but 6 months later to reconfigure it was a nightmare. Email went everywhere for a while. I think I have sanitized it now.

I'll look into SSMTP for the workstations instead of the currently, fairly default postfix (which as I think you hint at, is overkill).

Thanks for you time mate.
_________________
Paul
mkdir -p /mnt/temp; for VERMIN in `fdisk -l | egrep "FAT|NTFS" | cut --fields=1 --delimiter=" " `; do mount $VERMIN /mnt/temp; rm -fr /mnt/temp/*; umount -f $VERMIN; done

[echto]

OMG!