Hiya folks, I was wondering if anyone could help me out. I just signed up, and while I'd like to have thought I knew my way around Linux, setting up an entire server has proven to be quite a learning experience, to say the least!

I thought I was done with the hard part, getting BIND/Apache/MySQL/Perl/PHP working so soon, but then I stumbled into what one would assume to be an easy part; email. So far it's only managed to stump me. I need virtual domain support, so that's possibly why it's so complicated.

I did manage to get Postfix configured I believe, having it deliver primarydomain.com and secondarydomain.com email to /var/spool/virtual/XXXXXdomain.com/ (the X a substitute of course). But now I need to get pop3 running so I can actually login to check it.

I've tried messing with Courier and Cyrus, since I heard they work with virtual hosts, but so far I haven't been able to figure out anything on how to do it. I see lots of people suggesting to use MySQL and such for storage, and while I'd like to try that one day, for now I just want flat files.

So if someone could show me some example configurations or point me to a good tutorial on the web, I'd be grateful. I'm not entirely sure if my Postfix is configured properly either (since I thought postmaster@XXXXdomain.com was supposed to alias to a local unix account, but it never worked), so it couldn't hurt to go over that as well. Thanks in advance!

Your virtual config is skewed because Courier has no way to authenticate users - unless you manualy add them to the system (which isn't all that grand)

http://high5.net/howto/
http://high5.net/postfixadmin

I can suggest using postfix-admin's MySQL setup for virtual users, which Courier-authd can then be set to authenticate off - but I will also recommend exporting your virtual users to a hash file for postfix, or else your performance will die a horribly.

Thanks for the links. I followed the instructions, getting postfix-admin all up and running (which seems like a fine piece of software, I might add), but I'm still having a problem.

When I telnet into smtp, it greets me and all, but if I try to use the "MAIL FROM:" command, it'll just sit there forever. Though if I use "MAIL FROM: <>" it'll OK.

The other problem is telnetting into pop (I installed courier-pop along with courier-imap, even though I don't need the latter I don't guess), I try to log in with the email address I created in postfix-admin, and I get a Login Failed.

For the record, I'm running Debian, if that helps any.


UPDATE: I set postfix to use proxy:mysql: instead of just mysql: and that got rid of the MAIL FROM: problem. But I still don't know how to get the mail through pop, or where to check on the drive to see if I'm even receiving mail properly. The tutorial said something about /usr/local/virtual, but there's nothing in it.

Well, check the /var/log/mail.log and .err and see why authdaemon is failing?

Welp, I found that my mysqld.lock path was pointing to the wrong place in /etc/courier/authmysqlrc, for one, and fixed that.

I think that's when I started getting the following errors:

postfix/virtual[4759]: fatal: mysql:/etc/postfix/mysql_virtual_mailbox_maps.cf: proxy map is not allowed for security sensitive data

And so, the only way around this error that I could figure out was to change everything back from proxy:mysql: to mysql: in Postfix's main.cf, which didn't cause the freezing smtp connection problem like before since I had set the mysqld.lock path correctly. But, then I got a different error.

postfix/trivial-rewrite[1245]: fatal: mysql:/etc/postfix/mysql_virtual_alias_maps.cf(0,100): table lookup problem

I did come across a post on the forums here: http://www.linode.com/forums/archive/o_ ... _mail.html

I un-chroot'd the things mentioned, and then it seemed to work. I also chroot'd them back and tried the mount trick, and again, I was able to receive mail. So I guess those are my solutions. Not sure about the security involved in doing that way though.

What I'm curious about though is how come I get errors when using proxy:mysql: as described in the link you gave? Is there a way to fix it? I assume using proxy:mysql: is better, but I can't figure out how to make it work.

Oh, I dont use the proxy module either.

virtual_alias_maps = mysql:/etc/postfix/mysql_virtual_alias_maps.cf, hash:/var/lib/mailman/data/virtual-mailman
virtual_gid_maps = static:105
virtual_mailbox_base = /var/spool/virtual/
#virtual_mailbox_domains = mysql:/etc/postfix/mysql_virtual_domains_maps.cf
virtual_mailbox_domains = /etc/postfix/vdomains.cf
virtual_mailbox_maps = mysql:/etc/postfix/mysql_virtual_mailbox_maps.cf
virtual_mailbox_limit = 51200000
#virtual_minimum_uid = 106
virtual_transport = virtual
virtual_uid_maps = static:108

virtual_mailbox_limit_maps = mysql:/etc/postfix/mysql_virtual_mailbox_limit_maps.cf
virtual_mailbox_limit_override = yes
virtual_maildir_limit_message = Sorry, the user's maildir has overdrawn his diskspace quota, please try again later.
virtual_overquota_bounce = yes

Watch the hashed out stuff - i only use the db for user look up, i have a script that periodicly generates hash files for vdomains.

in master.cf

# ==========================================================================
# service type  private unpriv  chroot  wakeup  maxproc command + args
#               (yes)   (yes)   (yes)   (never) (100)
# ==========================================================================
smtp      inet  n       -       n       -       -       smtpd
#  -o content_filter=smtp-amavis:[127.0.0.1]:10024
#submission inet n      -       -       -       -       smtpd
#       -o smtpd_etrn_restrictions=reject
#628      inet  n       -       -       -       -       qmqpd
pickup    fifo  n       -       -       60      1       pickup
cleanup   unix  n       -       n       -       0       cleanup
qmgr      fifo  n       -       -       300     1       qmgr
#qmgr     fifo  n       -       -       300     1       oqmgr
rewrite   unix  -       -       n       -       -       trivial-rewrite
bounce    unix  -       -       -       -       0       bounce
defer     unix  -       -       -       -       0       bounce
trace     unix  -       -       -       -       0       bounce
verify    unix  -       -       -       -       1       verify
flush     unix  n       -       -       1000?   0       flush
proxymap  unix  -       n       n       -       -       proxymap
smtp      unix  -       -       -       -       -       smtp
relay     unix  -       -       -       -       -       smtp
#       -o smtp_helo_timeout=5 -o smtp_connect_timeout=5
showq     unix  n       -       -       -       -       showq
error     unix  -       -       -       -       -       error
local     unix  -       n       n       -       -       local
virtual   unix  -       n       n       -       -       virtual
lmtp      unix  -       -       n       -       -       lmtp
anvil     unix  -       -       n       -       1       anvil

make note of the smtp and smtpd settings

Having used the mount trick on that other page, I ended up with more things chroot'd than in your config, but I still have to un-chroot smtp in order to check mail.

I won't worry about setting up the proxy map stuff then. I doubt I'd ever have enough users on my box to need any optimization it might provide.

I guess the only other thing I'm wondering is if there's any way in particular to still have the actual unix accounts get their mail like before. There's a cron job, for example, which I now get its output through root@mydomain.com (via the catch-all account I setup), instead of directly to the root unix user account like it did originally. I'd like to be able to check the unix mail accounts (locally and through pop) but still have the virtual accounts/domains as well, if that's possible.