mydomain gets set to "localdomain"

[Alex R]

I'm on 16.04.1 with postfix 3.1.0-3. This is almost exactly the same bug that has been reported 8-12 years ago, explained in detail here: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=214741#10 ... which is exactly what I'm experiencing: $mydomain gets set to "localdomain" if my hostname is a single word (Ubuntu default) instead of a fqdn, unless I force mydomain=<domain.com> in main.cf

My host+mailname are configured as recommended by Ubuntu, and the postfix config is pretty default. The above upstream debian bug says it was fixed in v2.11, but I'm not sure ...

# postconf | grep ^mydomain
mydomain = localdomain

and then in turn:
# postconf | grep ^myhostname

Here are my configs:

# cat /etc/hosts
7x.xx.xxx.100 lin.foodomain.com lin
127.0.0.1 localhost loopback
127.0.1.1 ubuntu
::1 ip6-localhost ip6-loopback
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters

# cat /etc/hostname
lin

# cat /etc/mailname
foodomain.com

# hostname -f
lin.foodomain.com

# postconf -n
alias_database = hash:/etc/aliases
alias_maps = hash:/etc/aliases
append_dot_mydomain = no
biff = no
compatibility_level = 2
inet_interfaces = all
inet_protocols = ipv4
mailbox_size_limit = 0
masquerade_domains = $myorigin
milter_default_action = accept
milter_protocol = 2
mydestination = foodomain.com, lin.foodomain.com
mynetworks = 127.0.0.0/8 [::ffff:127.0.0.0]/104 [::1]/128
mynetworks_style = host
myorigin = /etc/mailname
non_smtpd_milters = $smtpd_milters
readme_directory = no
recipient_delimiter = +
relayhost =
smtp_tls_session_cache_database = btree:${data_directory}/smtp_scache
smtpd_banner = $myhostname ESMTP $mail_name (Ubuntu)
smtpd_milters = inet:localhost:8891
smtpd_relay_restrictions = permit_mynetworks permit_sasl_authenticated defer_unauth_destination
smtpd_sasl_auth_enable = yes
smtpd_tls_auth_only = yes
smtpd_tls_cert_file = /etc/ssl/certs/lin.foodomain.com.chained.crt
smtpd_tls_key_file = /etc/ssl/private/lin.foodomain.com.key
smtpd_tls_session_cache_database = btree:${data_directory}/smtpd_scache
smtpd_use_tls = yes
smtputf8_enable = no

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Ing. Pavel Milanes Costa (pavelmc)]

Hi, this bug still remains on bionic 18.04 and on the recently released 20.04

Is there any chance of a fix of this?

This bug is a long Confirmed one...

[Scott Kitterman (kitterman)]

I looked into the history of this. We used to have a patch in Debian to use gethostbyname to address this issue. It was dropped because it caused worse problems. Here's the relevant code snippet from the most recent postfix release:

    /*
     * If the local machine name is not in FQDN form, try to append the
     * contents of $mydomain. Use a default domain as a final workaround.
     *
     * DO NOT CALL GETHOSTBYNAME OR GETNAMEINFO HERE - IT MAKES EVERY POSTFIX
     * PROGRAM HANG WHEN DNS SERVICE IS UNAVAILABLE. IF YOU DON'T LIKE THE
     * DEFAULT, THEN EDIT MAIN.CF.
     */
    name = get_hostname();
    /* DO NOT CALL GETHOSTBYNAME OR GETNAMEINFO HERE - EDIT MAIN.CF */
    if ((dot = strchr(name, '.')) == 0) {
        /* DO NOT CALL GETHOSTBYNAME OR GETNAMEINFO HERE - EDIT MAIN.CF */
        if ((domain = mail_conf_lookup_eval(VAR_MYDOMAIN)) == 0)
            domain = DEF_MYDOMAIN;
        /* DO NOT CALL GETHOSTBYNAME OR GETNAMEINFO HERE - EDIT MAIN.CF */
        name = concatenate(name, ".", domain, (char *) 0);
    }
    /* DO NOT CALL GETHOSTBYNAME OR GETNAMEINFO HERE - EDIT MAIN.CF */
    return (name);

So the way to resolve this issue is to edit main.cf as people have indicated they are doing. "Fixing" this in postfix is not something we should do.