Resolved Postfix Overloads & Freezes Server

[tetrahall]

Hi,
Our Centos 7 server has been going down frequently. We've found that Postfix is the cause of the problem.

If Postfix service is stopped, the server doesn't freeze, and 'numproc' remains more or less steady within normal range.

When Postfix service is restarted, although no emails are sent, and (almost) none are received, the problem starts again. I noticed that the server appears to be sending emails to itself.

# mailq
-Queue ID- --Size-- ----Arrival Time---- -Sender/Recipient-------
7AF3C21E4F 718 Wed May 5 16:26:01 root root
001A421E4D 795 Wed May 5 17:01:03 root root
34EF621E2F 798 Wed May 5 16:01:05 root root
3344021E81 718 Wed May 5 17:26:01 root root
14D4A21E31 797 Wed May 5 18:01:05 root root

maillog (in /var/log/), displays few Postfix errors, such as these:-
May 5 18:19:41 domain postfix/cleanup[2712]: error: open /etc/postfix/body_checksmeta_directory: No such file or directory
May 5 18:19:41 domain postfix/cleanup[2712]: fatal: open dictionary: expecting "type:name" form instead of "="
May 5 18:19:42 domain postfix/master[2209]: warning: process /usr/libexec/postfix/cleanup pid 2712 exit status 1
May 5 18:19:42 domain postfix/master[2209]: warning: /usr/libexec/postfix/cleanup: bad command startup -- throttling

# service Postfix status

May 05 20:43:59 domain postfix/spawn[6904]: warning: /usr/lib64/plesk-9.0/postfix-srs: proc...ded
May 05 20:43:59 domain postfix/spawn[10512]: warning: /usr/lib64/plesk-9.0/postfix-srs: pro...ded
May 05 20:44:05 domain postfix/smtpd[7907]: fatal: unable to connect to the public cleanup ...ice
May 05 20:44:06 domain postfix/master[2209]: warning: process /usr/libexec/postfix/smtpd pi...s 1
May 05 20:44:06 domain postfix/master[2209]: warning: /usr/libexec/postfix/smtpd: bad comma...ing
May 05 20:44:07 domain postfix/cleanup[11479]: error: open /etc/postfix/body_checksmeta_dire...ry
May 05 20:44:07 domain postfix/cleanup[11479]: fatal: open dictionary: expecting "type:name"...="
May 05 20:44:08 domain postfix/master[2209]: warning: process /usr/libexec/postfix/cleanup ...s 1
May 05 20:44:08 domain postfix/master[2209]: warning: /usr/libexec/postfix/cleanup: bad com...ing

I was wondering if uninstalling Postfix, installing Qmail, and then reinstalling Postfix would remedy the problem with Postfix. If that is not the solution, I would appreciate your help in suggesting another one.

Regards

 

[tetrahall]

UPDATE:

Sending a simple test email now produces 500 internal server error.

In the error.log:
[Wed May 05 23:41:59.600529 2021] [fcgid:warn] [pid 1859:tid 139947980244736] [client 5.156.5.163:57246] mod_fcgid: read data timeout in 45 seconds,

[Wed May 05 23:41:59.600637 2021] [core:error] [pid 1859:tid 139947980244736] [client 5.156.5.163:57246] End of script output before headers: index.php,

I was wondering if "mod_fcgid: read data timeout in 45 seconds" can be fixed via some settings in Plesk ?

 

[weltonw]

It looks like the PHP scripts can't connect to the postfix server, and timesout

 

[tetrahall]

Thank you, john0001. Any idea how to go about solving this?

 

[weltonw]

A. See if postfix is actually running. If so, move to B
B. See if postfix is bound to the proper ports 465 + 587 and/or the sendmail binary works.
C. Check error logs in maillog for any errors, as well as syslog/messages
D. If it's sending messages to itself, there should be a the contents in /var/mail/root

 

[tetrahall]

Many thanks,john0001. I will work on that.
The webhost firm tell me Postfix uses port 25. However, both sending & receiving do not work.

 

[tetrahall]

Apparently the problem lies in file main.cf which needed a bit of tweaking.
It has now been resolved.