From: Bojan Zdrnja (no email)
Date: Wed Oct 01 2003 - 04:18:00 EDT
Hi,
First of all, sorry for a bit longer e-mail. But it was difficult to explain
the setup.
I have a (IMHO at least) little bit complex setup with virtual domains (and
amavisd - that shouldn't matter if I'm not wrong). Let me first explain
that. My server has to accept e-mails for 3 domains, same users on all of
them. The problem is with aliases (which are used for groups).
virtual_alias_domains = domain1, domain2, domain3
virtual_alias_maps = hash:/etc/postfix/virtual,
hash:/etc/postfix/virtaliases
virtual_mailbox_maps = hash:/etc/postfix/vmailbox
In /etc/postfix/virtual I have:
userid at domain1 userid at finaldomain
userid at domain2 userid at finaldomain
userid at domain3 userid at finaldomain
In /etc/postfix/vmailbox I defined:
userid at finaldomain @lmtp.host
This so far works ok - if I send an e-mail to that user, he receives his
e-mail on LMTP host.
Problem I have is with aliases. Those are defined in
/etc/postfix/virtaliases. The problem is that I have *a lot* of them (ie.
One alias group goes to like 4000 users, split among several aliases). This
is an example:
Group userid1,userid2,userid3.....userid400,group1
Group1 userid401,userid402,userid403....userid800,group2
Group2 ... And so on till lets say group8
Now, when I send an e-mail to lets say group at domain1 I need to have that
delivered to all users.
What happens is that e-mail gets delivered to exactly 1200 users (first 3
groups) and then it fails on fourth.
I also get one warning:
Oct 1 15:44:05 s_sys at host postfix/cleanup[10778]: warning: 3EEB233E76:
unreasonable virtual_alias_maps map expansion size for group at domain1
I get the following bounce:
Oct 1 15:44:14 s_sys at host postfix/error[11069]: 6339533E8C:
to=<userid560 at domain1>, orig_to=<group3 at domain1>, relay=none, delay=4,
status=bounced (user unknown in virtual alias table)
Notice that it says orig_to group3 in bounce message. I also get exactly 400
bounces - seems like it stops in group3 and doesn't process other aliases
(till group8).
Any pointers where to start searching the error? Group3 alias looks ok to
me, will go deeply through it.
Thanks,
Bojan
|
|
|