Hello Ubuntu Nation ... I used the "Virtual Users And Domains With Postfix, Courier, MySQL And SquirrelMail (Ubuntu 9.10)" HowTo to try to correct my postfix config when it suddenly stopped working. Now, I'm not doing virtual users or quotas, but everything else is in there. Here's the rundown of where things stand: What does work: Incoming mail to my users is handled properly Reading IMAP mail, either through Squirrelmail or an external client works Sending mail out through Squirrelmail works Note that Courier-IMAP is configured to authenticate through the MySQL DB, as is Apache for some directories. What does not work: Sending email through external clients times out on the connection attempt. This is configured (badly, it would seem) to authenticate to the same mysql database as Courier and Apache via saslauthd and the pam_mysql plugin. Note that I did have this working at one time. Unfortunately, an upgrade made it go wrong (I think), and since I couldn't see the problem (I seem to have been kept in the cache longer than any of the other users) I've been sitting on this for some time. Here's my main.cf: Code: soft_bounce = no biff = yesappend_dot_mydomain = yes readme_directory = /usr/share/doc/postfix data_directory = /var/lib/postfix mydomain = domain1.net myhostname = domain1.net alias_maps = hash:/etc/aliases alias_database = hash:/etc/aliases myorigin = /etc/mailname mydestination = domain1.net, domain2.net, domain1.com, domain2.com, domain1.org content_filter=smtp-amavis:[127.0.0.1]:10024 local_recipient_maps = unix:passwd.byname $alias_maps relay_domains = $mydestination relayhost = mynetworks_style = subnet mynetworks = 127.0.0.0/8 [::ffff:127.0.0.0]/104 [::1]/128 mailbox_size_limit = 1000000000 recipient_delimiter = + inet_interfaces = all unknown_local_recipient_reject_code = 550 unknown_address_reject_code = 554 unknown_hostname_reject_code = 554 unknown_client_reject_code = 554 home_mailbox = .Maildir/ disable_vrfy_command = yes strict_rfc821_envelopes = yes mailbox_command = /usr/bin/procmail -a "$EXTENSION" header_checks = regexp:/etc/postfix/header_checks smtpd_helo_required = yes smtpd_sender_restrictions = permit_mynetworks, permit_sasl_authenticated smtpd_recipient_restrictions = permit_mynetworks, permit_sasl_authenticated, reject_unauth_destination, reject_non_fqdn_sender, reject_non_fqdn_recipient tls_random_source = dev:/dev/urandom smtpd_use_tls=yes smtpd_tls_received_header = yes smtpd_tls_session_cache_timeout = 3600s smtpd_tls_session_cache_database = btree:${data_directory}/smtpd_scache smtpd_tls_key_file = /etc/ssl/courier-imap/cert.net.key smtpd_tls_cert_file = /etc/ssl/courier-imap/cert.net.crt smtpd_tls_CAfile = /etc/ssl/courier-imap/cert.crt smtpd_tls_loglevel = 3 smtp_use_tls = yes smtp_tls_loglevel = 3 smtp_tls_note_starttls_offer = yes smtpd_sasl_type = cyrus smtpd_sasl_auth_enable = yes smtpd_sasl_local_domain = broken_sasl_auth_clients = yes smtpd_sasl_path = smtpd smtpd_sasl_authenticated_header = yes smtpd_sasl_security_options = noanonymous html_directory = /usr/share/doc/postfix/html And my /usr/lib/sasl2/smtpd.conf: Code: pwcheck_method: saslauthd mech_list: plain login cram-md5 digest-md5 allow_plaintext: true auxprop_plugin: mysql sql_hostnames: 127.0.0.1 sql_user: mailuser sql_passwd: mypasswd sql_database: passwords sql_select: select clear from passwd where id = '%u' This is symlinked to /etc/postfix/sasl/smtpd.conf I know the relayhost is empty. This is intentional, and does not seem to be causing issues with mail sent via Squirrelmail. The problem is that when trying to connect to postfix with a mail client like Thunderbird, Apple Mail, Outlook, etc. the connection times out. Same with telnetting in to port 25. I have verified that port 25 is open in the firewall and that the port is listening. When trying to connect, I can even see the SYN_RECV state on the server, but nothing whatsoever shows in any of the logfiles. When I telnet in to port 25 from the server, I see this: Code: # telnet localhost 25 Trying 127.0.0.1... Connected to localhost. Escape character is '^]'. 220 domain1.net ESMTP Postfix ehlo localhost 250-keyslapper.net 250-PIPELINING 250-SIZE 10240000 250-ETRN 250-STARTTLS 250-AUTH CRAM-MD5 LOGIN DIGEST-MD5 PLAIN 250-AUTH=CRAM-MD5 LOGIN DIGEST-MD5 PLAIN 250-ENHANCEDSTATUSCODES 250-8BITMIME 250 DSN When I telnet in from an outside host, I get a timeout. I wouldn't be so confused if I weren't getting the SYN_RECV on the host. I'm pretty much reaching the end of my rope with this, so any suggestions on how to resolve this would be greatly appreciated.
Hi Falko, netstat -tap is showing the master process listening on port 25, and there was nothing whatsoever in the mail logs. There were definitely packets coming through, but that only served to obscure the problem. I've just come across a solution that I probably should have though of before: Turns out port 25 was being blocked by my ISP - why it was still letting some packets through, I have no idea. I had gotten into the habit of using webmail when I changed my home ISP and didn't notice the problem until I was past the association. My other users didn't notice it at the same time because my brother was traveling and connecting through hotspots that were not blocking that port. The fix was to uncomment the submission configuration in master.cf and open port 587 in the firewall. Other things I had to remember: * Open the submission port in the firewall. * Update the Squirrelmail configuration to use the correct port. * Set the -o smtpd_tls_security_level=may in the master.cf submission config to allow Squirrelmail to connect without TLS, as well as external clients that may not be able to use it. * Add the following to the master.cf submission: -o mynetworks=127.0.0.0/8 * And change the smtpd_client_restrictions setting to include mynetworks. Without these changes, I'd have had to jump through hoops to get Squirrelmail to go through the SMTP authentication as well as the IMAP authentication. Didn't seem worthwhile. So, now I'm on to the next problem. It seems the postfix - sasl - pam - mysql authentication chain is not quite right. But at least I'm actually getting useful information there. If I can't find the solution to this one, I'll post another thread. I hope this helps someone else avoid the frustration I went through. Thank you!
Cyrus and postfix Hi, sorry if I have used the wrong thread to ask the question. but I thought it's some what related. apparently, I have a fedora running cyrus and postifx to host several domains with similar setting with virutal users, domains with postifx, mysql and squirrelmail. I have also a ubuntu 9.04 running ispconifg 2. I have just moved one domain say xxx.com from the fedora to ubuntu. Now, the problem is emails sent from the fedora to the xxx.com (now docked in ubuntu). when investigating the maillog in fedora, the xxx.com email address resolve back to the fedora host name which is the old host. how do I get it to point to the new host? any help appreciated
You'll need to change the DNS pointers to direct your domain to the new host. You can also trick the old host into recognizing the new host by putting the IP of the new host in the /etc/hosts file on the old host. It will look something like this: 123.45.67.89 xxx.com HTH