Loading ...
Sorry, an error occurred while loading the content.

292065Re: LDAP canonical_maps and domain rewriting

Expand Messages
  • Patrick Lists
    Mar 19, 2013
      Hi Fernando,

      On 03/19/2013 01:02 PM, Fernando Maior wrote:
      > Hello,
      > All this seems to be something very different from what postfix and
      > other smtp usually does. So, may be the problem is with the concept, not
      > with the implementation.
      > May I ask you why you need to change the domain name part of the mail
      > delivery address? Can you provide us with information on your mail
      > accepting and delivery needs?
      > May be if you look from a different direction, you can see a different -
      > and more appropriate - sollution.

      I don't think I'm doing something out of the ordinary but that's just me
      :-) Here's it goes:

      I use unique email addresses (aliases) for every website I register or
      where I order something. Right now I have close to 300 aliases using
      several different domains (private & business). On my current ancient
      CentOS5 mailserver Postfix handles those domains and the aliases. So all
      mail is processed by postfix and then delivered to dovecot. The new
      mailserver will use Postfix plus some groupware software and the concept
      is taken from http://www.postfix.org/VIRTUAL_README.html: Non-Postfix
      mailbox store: separate domains, non-UNIX accounts.

      So I'm using virtual_mailbox_domains, virtual_maibox_maps,
      virtual_alias_maps, virtual_transport and canonical_maps and the
      accounts are stored in OpenLDAP.

      Examples of how email addresses are handled:

      amazon@... is delivered to myaccount@... because
      amazon@... is an alias of myaccount@....

      biz@... is rewritten to biz@... because it's in
      canonical_maps and then delivered to myaccount@... because
      biz@... is an alias of myaccount@....

      The second example is the reason why I asked about canonical_maps with
      LDAP that would do @... -> @....

      In my new test setup this all works fine although I don't doubt that
      Victor could find something odd in my setup that requires me to read
      many more RFCs to get a clue :-)

      Hope this makes sense.

    • Show all 18 messages in this topic