I have three mail server questions, if I might. 1. I have set up most of the systems on my network to use my mail server for internal mail, things like ossec messages, apticron, etc. Most hosts run exim, so I have them set up with the mail server as my smarthost. However, one box, running Proxmox VE, is running postfix. I configured it via dpkg-reconfigure, told it the machine was a satellite host, then pointed it to the smarthost. I also updated /etc/aliases to point all mails on the local machine to my domain email (i.e. postmaster, nobody, hostmaster, webmaster and www all redirect to root, then root points to [email protected]). However, when I send email from the host, it goes to the mail server, but then tries to go back to the host, and I don't understand why. I updated the aliases file and also ran newaliases.
On the host: Feb 24 13:16:05 hornet postfix/smtp[5703]: A322119E872: to=<[email protected]>, relay=mail.bar.com[192.168.1.12]:25, delay=0.2, delays=0.02/0.01/0.01/0.17, dsn=2.0.0, status=sent (250 2.0.0 Ok: queued as AA87EF1851) On the mail server: Feb 24 13:16:06 merrimac postfix/qmgr[14271]: A5D97F1852: from=< [email protected]>, size=1143, nrcpt=1 (queue active) Feb 24 13:16:06 merrimac postfix/smtp[16354]: connect to hornet.bar.com[192.168.1.13]:25: Connection refused Feb 24 13:16:07 merrimac postfix/smtp[16354]: E0574F1862: to=< [email protected]>, relay=none, delay=0.05, delays=0.04/0.01/0/0, dsn=4.4.1, status=deferred (connect to hornet.bar.com[192.168.1.13]:25: Connection refused) I know I could open up hornet's postfix to the entire address space (it's set to the default -- 127.0.0.0/8), but it seems to me that would open up a nasty email loop. 2. Currently, I am running Zimbra on a P4 box. I would really *really* like to get off of Zimbra. First of all, the Debian packages are deprecated and doubt, this being the case, that they will come out with new packages for squeeze. Secondly, my load average hovers between 4 and 9 on this because of all the java process that seem so rampant on zimbra. So I am setting up a VM with postfix, dovecot and (optionally) roundcube. However, I am having issues with getting dovecot working. I set it up, but I am unable to log in. set up dovecot, I'm able to telnet to ports 143 and 993 as well as telnetting using -starttls on 143, and they all work. However, when I do a mutt -f imap://user@<host>/user either from the mail machine or my workstation across the network, I get authentication failed, and it blows me out of mutt. The logs on the mail machine show: Feb 24 18:20:14 uhura dovecot: imap-login: Aborted login (auth failed, 2 attempts): user=<user>, method=PLAIN, rip=192.168.1.1, lip=192.168.1.53, TLS Feb 24 18:21:55 uhura dovecot: imap-login: Aborted login (auth failed, 2 attempts): user=<user>, method=PLAIN, rip=192.168.1.53, lip=192.168.1.53, TLS I used the guide at http://www.splitbrain.org/blog/2008-01/31-setup_postfix_and_dovecot_on_debian_etch. I know its for etch, but I was hoping it would work for wheezy/sid. Does anyone have a good guide, or from the limited information I can find, see where I am missing something? 3. Does anyone have a roundcube guide for Debian? Thanks, --b
