• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Resolved plesk update broke postfix; created broken/improper postfix master.cf

TomBoB

Regular Pleskian
Hi all,

on CentOS 7.3, Plesk 17.5 #19

after the latest Plesk update, postfix stopped working properly. Clients coming from certain specific ISPs and wanting to log in to send their mail would be greeted with either an error message
"None of the authentication methods supported by this client are supported by your server" or
"too many connections from [public ip address of ISP]".

Clients have to use submission port 587 with authentication enabled.

The maillog has lots of entries:
Aug 23 22:23:50 rhino postfix/smtpd[16139]: connect from unknown[168.167.85.38]
Aug 23 22:23:50 rhino postfix/smtpd[17812]: connect from unknown[41.223.195.3]
Aug 23 22:23:52 rhino postfix/smtpd[16139]: lost connection after EHLO from unknown[168.167.85.38]
Aug 23 22:23:52 rhino postfix/smtpd[16139]: disconnect from unknown[168.167.85.38]
Aug 23 22:23:52 rhino postfix/smtpd[16788]: connect from unknown[168.167.85.38]
Aug 23 22:23:55 rhino postfix/smtpd[16788]: SSL_accept error from unknown[168.167.85.38]: lost connection
Aug 23 22:23:55 rhino postfix/smtpd[16788]: lost connection after STARTTLS from unknown[168.167.85.38]

A quick investigation revealed that the Plesk update somehow added an additional bind line to the master.cf - apparently ignoring / not recognizing the proper existing one.

Although an ipv6 address is configured for the server, I was under the impression that no domains were using it. A check revealed that 3 domains actually are configured to use the ipv6 address.

The oiginal master.cf:
...
plesk_saslauthd unix y y n - 1 plesk_saslauthd status=5 listen=6 dbpath=/var/spool/postfix/plesk/passwd.db

plesk-62.1.1.1- unix - n n - - smtp -o smtp_bind_address=62.1.1.1 -o smtp_bind_address6= -o smtp_address_preference=ipv4

submission inet n - n - - smtpd -o smtpd_enforce_tls=yes -o smtpd_tls_security_level=encrypt -o smtpd_sasl_auth_enable=yes -o smtpd_client_restrictions=permit_sasl_authenticated,reject -o smtpd_sender_restrictions= -o smtpd_recipient_restrictions=permit_mynetworks,permit_sasl_authenticated,reject_unauth_destination

whereas the master.cf after the update:
...
plesk_saslauthd unix y y n - 1 plesk_saslauthd status=5 listen=6 dbpath=/var/spool/postfix/plesk/passwd.db

plesk-62.1.1.1- unix - n n - - smtp -o smtp_bind_address=62.1.1.1 -o smtp_bind_address6= -o smtp_address_preference=ipv4

submission inet n - n - - smtpd -o smtpd_enforce_tls=yes -o smtpd_tls_security_level=encrypt -o smtpd_sasl_auth_enable=yes -o smtpd_client_restrictions=permit_sasl_authenticated,reject -o smtpd_sender_restrictions= -o smtpd_recipient_restrictions=permit_mynetworks,permit_sasl_authenticated,reject_unauth_destination

plesk-62.1.1.1-2001-4178-5-111-62-1-1-1 unix - n n - - smtp -o smtp_bind_address=62.1.1.1 -o smtp_bind_address6=2001:4178:5:111:62:1:1:1 -o smtp_address_preference=ipv4

Did anyone come across this scenario?
Can anyone explain? Don't have enough background knowledge (yet) to fully grasp the issue...

Removing the last/incorrect/additonal bind config and restarting postfix resolved the issue of postfix not accepting connections.

Thanks, Tom
 
Back
Top